The GLICO Model for Glioblastoma Multiforme by

Glioblastoma Multiforme

kathleen_bialik — 2019-11-20 16:17:39 UTC

Glioblastoma Multiforme (GBM) are intrinsic brain tumours that are believed to originate from neuroglial stem cells.¹
One of the most deadly forms of cancer, afflicted individuals have a 5 year survival rate of <5%.¹
Glioblastomas are mitotically active tumors characterized by microvascular proliferation and necrosis, creation of their own blood supply, and the ability to active tumour associated astrocytes.⁵
Due to the complexity of the disease and the cerebral environment that the tumours reside, GBM is extremely difficult to study.⁴

Glioma Stem Cells (GSCs)

kathleen_bialik — 2019-11-20 16:18:27 UTC

Generally, stem cells (SC) have the capacity to self renew and differentiate into more specific types of cells.³
Referred to as “potency”, the amount of options that are available to differentiate into relies heavily on the origin of the cell and well as the microenvironment it resides in.²
GSCs are pluripotent stem cells that reside in cerebral microenvironments, promote the growth of GBM tumours, and have exhibited therapeutic resistance to current forms of treatment.⁶
GSCs exhibit resistance to ionizing radiation (damaging the DNA of cancerous tissue leading to cellular death) and cytotoxic drugs (chemically killing the cells directly).³
The protective ability that these cells have on the tumor make them ideal candidates for studying the mechanistic neuroprotection and eventually defining a treatment that can remove these GSCs from the equation.³

Issues With Current Models

kathleen_bialik — 2019-11-20 16:18:46 UTC

In vitro experimental methods for GBM consist of cell autonomous interactions via a 2D cell culture.¹
The concept of cell autonomy is not viable for accurate representation of GBM, due to the GSC’s reliance on the cerebral microenvironment as well as the documented cell to cell interactions that characterize tumors.¹
Theoretically 2D co-culturing of many different types of brain cells along with GSCs to study cell to cell interactions is an option, but the lack of the microenvironment will result in unnatural differentiation of the SCs.¹
- Lack of the tumour microenvironment in 2D models also doesn't properly model the effects of chemotherapeutic drugs in patients.
As for in vivo studies like mouse models, these can be very lengthy and very expensive, requiring model organisms and ample time for tumor formation and destruction. There is a need for a new model that can accurately resemble the environment in which GBM thrives, leading to the introduction of an ex vivo approach.¹

Organoid Formation

kathleen_bialik — 2019-11-20 16:20:03 UTC

The organoid is formed beginning with the culturing of human pluripotent stem cells (hPSCs) in media.⁷
After 6 days of incubation, embryoid bodies begin to form, and the neuroectoderm is removed from these and sub cultured until Day 11 on neural induction media.⁷
Following this incubation, the neuroectoderm is placed onto a matrigel droplet into differentiation media (contains stem cell differentiation factors and signals) to expand into the neuroepithelium.⁷
On Day 15, the neuroepithelium is transferred into a spinning bioreactor for formation of the cerebral tissue.⁷

Goals of the GLICO Model

kathleen_bialik — 2019-11-20 16:20:32 UTC

Due to the complexity of GBM tumours and the tumour microenvironment, the main question we can ask is how can the researchers refine technology using stem cell derived models in order to study the basic mechanisms for the effects that the glioma stem cells and the microenvironment have on glioblastomas?¹

Cerebral Organoid Synthesis

kathleen_bialik — 2019-11-20 16:34:13 UTC

Process of synthesizing a cerebral organoid beginning from the Day 0 culture of human Pluripotent Stem Cells.
Lancaster et. al. Nature. 2013.

References

kathleen_bialik — 2019-11-20 17:00:07 UTC

Linkous A. et. al. Modeling Patient-Derived Glioblastoma with Cerebral Organoids. Cell Reports. 2019; 26(12): 3203-3211.
Plath K, Lowry W. Progress in understanding reprogramming to the induced pluripotent state. Nature Reviews Genetic. 2011;12(4):253–265.
Aderetti D, Hira V, Molenaar R, van Noorden C. The hypoxic peri-arteriolar glioma stem cell niche, an integrated concept of five types of niches in human glioblastoma. Biochimica et Biophysica Acta- Reviews on Cancer. 2018; 1869(2): 346-354.
Le Rhun E, Preusser M, Roth P, Reardon D, van den Bent M, Wen P, Reifenberger G, Weller M. Molecular targeted therapy of glioblastoma. Cancer Treatment Reviews. 2019; 80: 101896.
Szeliga M, Albrecht J. Opposing roles of glutaminase isoforms in determining glioblastoma cell phenotype. Neurochemistry International. 2015; 88: 6-9.
Tasaki T, Jujita M, Okuda T, Yoneshige A, Nakata S, Yamashita K, Yoshioka H, Izumoto S, Kato A. MET Expressed in Glioma Stem Cells Is a Potent Therapeutic Target for Glioblastoma Multiforme. International Journal of Cancer Research and Treatment. 2016; 36(7): 3571-3577.
Sawai T, Sakaguchi H, Thomas E, Takahashi J, Fujita M. The ethics of cerebral organoid research: being conscious of consciousness. ISSCR. 2019; 13: 440-447.
Osswald M., Jung E., Sahm F., Solecki G., Venkataramani V., Blaes J., Weil S., Horstmann,H., Wiestler B., Syed M., et al. (2015). Brain tumour cells interconnect to a functional and resistant network. Nature 528, 93–98.
Bian S, Repic M, Guo Z, Kavirayani A, Burkard T, Bagley JA, Krauditsch C, Knoblich JA. Genetically engineered cerebral organoids model brain tumour formation. Nat Methods. 2018; 15: 631-639.
Krieger TG, Tirier SM, Park T, Eisenmann T, Peterziel H, Angel P, Eils P, Conrad C. Modeling glioblastoma invasion using human brain organoids and single-cell transcriptomics. bioRxiv. 2019:1-13.

Original Publication

rifatsajid96 — 2019-11-29 04:34:57 UTC

The GLICO Model

rifatsajid96 — 2019-11-29 04:41:18 UTC

Using a GBM Cerebral Organoid (GLICO) allows human GBM ex vivo modelling for high throughput drug screening.¹
This method co-cultures GSCs with cerebral organoids derived from human embryonic stem cells (hESCs) to resemble the microenvironment in which the tumour cells thrive in in human patients.¹
The introduction of pluripotent stem cells allows for the vast differentiation into all parts of the brain due to cell to cell interactions as well as interactions with their microenvironments.³
Using sub cultured hESCs, propagation of the neuroectoderm into a spinning bioreactor allows for the precise environment needed to form the cerebrum as an organoid.⁷

Methods & Results

rifatsajid96 — 2019-11-29 04:45:16 UTC

Patient-Derived GSCs Form Infiltrative Tumors in Cerebral Organoids (GLICO)

rifatsajid96 — 2019-11-29 04:48:26 UTC

To isolate GSCs, GBM tumors were surgically removed, washed, and enzymatically dissociated into single cells.
These cells were then cultured in medium that promotes the growth of GSCs.
Patient-derived GSCs were then co-cultured with cerebral organoids for 24 hours, after which the GSCs infiltrated the cerebral organoid.
In the figure on the left, the authors labelled GSCs with GFP using lentivirus prior to co-culturing, this allowed them to visualize the GSCs after infiltration.
- Considerable tumor growth was detected 1 week after co-culture.
On the right is H&E staining showing the bulk tumor, which corresponds to the GFP+ cells in the inset image, as well as normal organoid tissue.

Strengths of the GLICO Model + Significance of Findings

natalie_uzynski — 2019-11-29 04:51:03 UTC

The model advanced GBM research in a human brain environment by overcoming many of the limitations that past models of disease had - holds significance because gaining a greater and accurate understanding of the disease will further contribute to the discovery of effective treatments
The model addresses the critical issue of cell to cell interactions and allows the study of patient-specific GBMs within a similar microenvironment to that of a primitive human brain ¹ - significant because past GBM models did not (ex. tumour organoids, 2D models, and patient-derived mouse xenograft models, all previous model of GBM). This is also significant when testing disease treatments because sometimes treatments can be effective in non-human models but have no effect when introduced in humans, this model provides a more replicative environment of the human condition.
The model is amenable to experimental manipulation, drug treatment, and controlled physiological and environmental variables (possible because the model is grown ex vivo, advantages over an in vivo approach)¹. This is significant for future directions to enhance the model (ex. provide blood brain barrier characteristics).
The model is scalable; allows the generation of hundreds of patient-specific GLICOs for high throughput screening. Significant because it is a novel approach with high potential, it is scalable in ways not possible by any previous in vivo models.
The methodology in the paper was quite effective as the authors managed to address many key features of GBM into their GLICO model. This includes; ensuring the the tumours within the GLICO maintained the genetic signalling networks found in actual patient-derived tumours by surveilling for EGFR which is commonly up regulated in tumour cells (Figure 2D), or by evaluating their GLICO model to ensure that It mirrored the aggressive nature of GBM disease pathology in humans.

Infiltrating GSCs Destroy Cerebral Organoid Tissue

rifatsajid96 — 2019-11-29 04:52:39 UTC

It is well known that GBM disease pathology in humans is very aggressive, so the authors wanted to evaluate if their GLICO model mirrored the aggressive pathology in vitro.
- Again, H&E staining was used.
The top half of this figure shows a control organoid that was not co-cultured with GSCs.
- The tissue and structures of the organoid are still intact.
The bottom half of this figure is a GLICO that was co-cultured with GSCs. The “923” refers to the GSC cell line used, as there are multiple GSC cell lines.
- Here, there is a destruction of organoid tissue as the invasion and proliferation of GSCs induced abundant necrosis, which is shown by the black arrow. The blue arrow is pointing to the tumor cells themselves.

2D GSC Culture and GLICOs Differ in Response to Chemotherapeutic Agents

rifatsajid96 — 2019-11-29 04:57:12 UTC

Given that the GLICO model well recapitulated certain characteristics of human GMB tumours, the authors next wanted to determine the response of GLICOs to chemotherapeutic agents.
As previously mentioned, one limitation of the 2D GSC model is that the in vivo tumor microenvironment is not well recapitulated.
This environment can have profound effects on the ability of tumor cells to respond to genotoxic damage from chemotherapy.
So, the authors wanted to determine if the GLICO model was a better model for the effects themicroenvironment has during chemotherapy.
They tested common GBM chemotherapy drugs temozolomide (TMZ) and bis-chloroethylnitrosourea (BCNU) on both 2D GSC cultures and GLICOs in a cytotoxicity assay.
- This assay measures cell death after chemical treatment by evaluating how many cells take up trypan blue dye.
- Live cells have intact cell membranes, so they will not take up the dye. However, dead cells have ruptured cell membranes and will take up the dye.
This figure is showing the effect of these drugs on two different 2D GSC cultures, 827 and 923; TMZ is on the left, and BCNU is on the right.
For both 2D GSC cultures, there was a dose-dependent decrease in cell viability observed by treatment with both drugs.
- So, the higher the dose of the drug, the lower the cell number.

2D GSC Culture and GLICOs Differ in Response to Chemotherapeutic Agents (Continued)

rifatsajid96 — 2019-11-29 05:00:37 UTC

The dramatic decrease seen in 2D GSC cultures was not seen in the GLICOs.
Untreated cells from the same GSC lines from before were co-cultured with organoids.
- These GSCs were tagged with luciferase, an enzyme that produces bioluminescence, instead of GFP because it allows more accurate measurements of fluorescence.
In this figure, we are looking at changes in luciferase activity after drug treatment,as GSCs killed in this assay no longer express bioluminescence.
In the top half of the figureare GLICOs with GSCs from line 827.
- TMZ treatment only resulted in a 23% reduction in tumor growth, while BCNU treatment resulted in 91% reduction.
- The reduction in tumor growth is reflected in the decreased bioluminescence.
In the GLICOs with GSCs from line 923 in the bottom half of this figure;
- TMZ treatment resulted in a 43% reduction in tumor growth, while BCNU treatment resulted in only a 5% reduction.
- Again, the reduction in tumor growth is reflected in the decreased bioluminescence.

Weaknesses of the GLICO Model + Paper

natalie_uzynski — 2019-11-29 05:02:01 UTC

The model will never be completed, as an exact representative of the human clinical disease cannot be obtained.¹
The paper presents many unclear results and unsubstantial evidence is used to support the results:
- Ex. Figure 3C which quantifies EGFR copy number variation in 2D vs. GLICO samples from patients using EGFR amplification. The paper fails to explain the why the results are what they are. The authors briefly mention that the 0810 strain maintains high levels of EGFR expression (lost in 0607 GSCs), but do not justify why.
  - It is also unclear when the authors state that the rapid loss of EGFR amplification once GBM cells are cultured in 2D is well described, because 0810 does not seem to show a loss in EGFR amplification and the authors never expanded upon why this is so
  - "0810 2D" is the only value with no statistical significance, this would suggest that the decrease in 0810 2D is not significantly different from the '0810 GLICO'.

GLICOs Preserve Key Genetic Changes of the Parental Tumor

rifatsajid96 — 2019-11-29 05:03:57 UTC

In addition to the ability of the GLICO model to recapitulate the pathological phenotype of human GBM, the authors wanted to investigate if the tumors within the GLICO maintain the genetic signaling networks found in actual tumors from patients.
In GBM patients, epidermal growth factor receptor (EGFR) is commonly upregulated in tumor cells.
In this figure, we are looking at the number of EGFR copies in the 2D GSC model, GSCs in the GLICO, and non-tumor cells from the same GLICO (these are normal organoid cells).
- Copy number was measured through quantitative PCR, which monitors the number of EGFR mRNA transcripts over time.
The left hand side is showing the 2D GSC cultures and GLICOs derived from GSC line 0607, and on the right is the same but with GSC line 0810.
- Every dot represents the number of EGFR copies from independent replicates, while the black bar represents the average copy number among the replicates.
If we take a look at the left side of the figure, we see that GSCs in the GLICO have higher EGFR copy numbers compared to 2D GSC cultures as well as normal GLICO cells.
On the right side, GSCs in the GLICO as well as those in the 2D model have higher EGFR copy numbers than the normal GLICO cells.
Indeed, GLICOs do preserve genetic changes of the parental tumor.

Comparison of the GLICO Model to Similar Publications

natalie_uzynski — 2019-11-29 05:52:58 UTC

Different (but similar) approaches to studying GBM :

1. Bian et al. established a 3D in vitro similar to the GLICO model. Their model was named the neoplastic cerebral organoid (neoCOR) and it differed from the GLICO model as it used CRISPR/Cas9-mediated mutagenesis to introduce oncogenic mutations in cerebral organoids to recapitulate brain tumour genesis. The neoCOR and GLICO models are both effective in recapitulating the 3D tumour environment, and they are both a valuable tool for testing potential drugs in a personalized setting for GBM.

2. Krieger et al. also proposed an experimental approach to study GBM and also normal brain cells of the neuronal lineage in vitro.This approach differed from the GLICO model because the authors used induced pluripotent stem cell-derived human cerebral organoids as a 3D scaffold for invasion of patient-derived GBM cells and used a combination of techniques (tissue clearing, confocal microscopy etc.) to analyze tumour microbe development. Their approach suggested potential therapeutic targets through the examination of interactions between tumour and normal brain cells.

Concluding Remark

natalie_uzynski — 2019-11-29 05:53:41 UTC

The authors managed to create a model (GLICO) that effectively mirrored GBM disease pathology in humans, therefore making it an effective model to study and test potential treatments on.
Apart from there being ethical concerns in regards to the GLICO model, and further research required to fully understanding it and its potential, the model has ultimately demonstrated promising advances to studying GBM in a novel and effective manner over past models. The evidence effectively demonstrated what a powerful tool the GLICO model is for modelling therapeutic interventions and for studying GBM biology in a primitive human brain environment.

Ethical Challenges

natalie_uzynski — 2019-11-29 06:00:19 UTC

The GLICO model uses human embryonic stem cell (hESC)- derived cerebral organoids. The use of embryonic stem cells is very controversial. Many researchers object to the practice of growing human embryos in a laboratory.
- Some researchers even argue that cerebral organoids themselves have some sort of mental activity such as cognition, and so this questions if they currently, or in the future will have consciousness.¹⁰
- An alternate approach would be to study GBM using induced pluripotent stem cells as they are less controversial but still effective.

GSCs Proliferate Within GLICO

rifatsajid96 — 2019-11-29 19:06:09 UTC

Human GBM tumours infiltrate the brain through proliferation.
To confirm that a portion of the glioma cells within the GLICO were indeed actively proliferating, GLICOs were pulsed with EdU.
- EdU is a modified thymidine analog that is taken up by proliferating cells, and can be fluorescently labelled.
In this figure, proliferating GSCs take up EdU and are thus double positive for both GFP and EdU. Whereas quiescent GSCs are only GFP+.
Thus the authors confirmed that not only are the GFP-tagged GSCs infiltrating the GLICO, but they are also proliferating within the organoid.

Future Directions with the GLICO Model

natalie_uzynski — 2019-11-30 00:33:11 UTC

Enhancing the model to reach its full potential in effectively modelling GBM while also overcoming some of the ethical concerns. Doing so by taking advantage of the model's amenability to bioengineering methods, by using novel bioengineering approaches to introduce both a perfused vasculature (with blood brain barrier characteristics), and an immunologic niche (by taking advantage of the autologous nature of the model)¹

Addressing some of the unknowns
- Determining whether the increased infiltrative pattern in older cerebral organoids is associated with the even more aggressive natural history of GBM in elderly patients (in comparison to younger patients).
I suggest computing drug sensitivity comparisons in GLICO with in vivo tumour drug sensitivities in order to determine if the model is actually able to improve the predictive efficiency of in vitro and ex vivo therapeutic screens (as this is not yet confirmed).
Moving on to different model organism systems to study the GLICO model (ex. mice), with hopes of developing an effective treatment for GBM-patients in the future

Successful Data, Successful Model

natalie_uzynski — 2019-11-30 00:58:40 UTC

The goal of the project was met as the authors managed to successfully advance the technology used for research revolving glioblastoma to overcome limitations of current preclinical GBM models, through the establishment of a model system in which they could retro-engineer patient- specific GBMs using patient-derived glioma stem cells (GSCs) and human embryonic stem cell derived cerebral organoids.¹

Figures 3A exemplify effectiveness as the isogenic GSC lines are more resistant to chemotherapeutic drugs when grown within the microenvironment of the GLICO than when grown under 2D conditions.¹

(Other in vitro models do not really have this tumour microenvironment and this is a major limitation to them as the lack of a “normal” human microenvironment makes many tumour cell lines unable to accurately reproduce GBM biology, therefore making them non-true representations of the human brain).

Further results also demonstrated the success of the model. The authors used an effective approach of a cytotoxicity assay to observe the effects of GBM chemotherapeutic drugs (TMZ and BCNU) have on 2D GSC cultures versus GLICO. Their findings demonstrated that in presence of a higher dose, there is lower cell count indicating an effective treatment.
- results showed that the model 1) accurately modelled GBM and 2) successfully targeted with these common therapeutic drugs

Figure 3A

natalie_uzynski — 2019-11-30 01:06:42 UTC

Tumour Growth in GLICOs Is Supported by a Network of Microtubes

rifatsajid96 — 2019-11-30 02:28:11 UTC

It was previously reported by Oswald et al. (2015) that human GBM tumours in situ express a network of interconnecting microtubes.
- This network is thought to facilitate the proliferation and invasion of the tumor through communication between tumour cells.
The authors of this study wanted to determine if GLICOs could also model this process observed during human GBM tumour growth.
Figure A shows microtubes connecting different GSC colonies in 2D culture, as indicated by the black arrows.
These microtubes were also found observed between GFP-tagged GSCs in the GLICO, shown in Figure B with white arrows.

rifatsajid96 — 2019-11-30 05:47:38 UTC

In this study, Linkous et al. addressed limitations of current models of GBM by proposing their model named the GLICO model. The model combined patient-specific GBMs using patient-derived glioma stem cells and human cerebral organoids. The resulting products were tumours that closely phenocopied patient GBMs due to their similarity in microenvironments, thereby making them a proprietary and effective approach to studying GBM in hopes of developing a treatment for the disease.

Figure 3C

natalie_uzynski — 2019-11-30 20:50:36 UTC

Powerpoint Presentation

kathleen_bialik — 2019-11-30 22:06:31 UTC

Discussion/Conclusions

rifatsajid96 — 2019-11-30 23:27:57 UTC

Introduction

rifatsajid96 — 2019-12-01 01:59:28 UTC

Making Cerebral Organoids

rifatsajid96 — 2019-12-01 02:37:29 UTC