High glucose concentrations mask cellular phenotypes in a stem cell model of tuberous sclerosis complex

Tuberous sclerosis complex (TSC) is a neurodevelopmental disorder caused by deletions in the TSC1 or TSC2 genes that is associated with epilepsy in up to 90% of patients. Seizures are suggested to start in benign brain tumors, cortical tubers, or in the perituberal tissue making these tubers an interesting target for further research into mechanisms underlying epileptogenesis in TSC. Animal models of TSC insuf ﬁ ciently capture the neurodevelopmental biology of cortical tubers, and hence, human stem cell-based in vitro models of TSC are being increasinglyexplored inattemptsto recapitulatetuber developmentandepileptogenesisinTSC. However, in vitro culture conditions for stem cell-derived neurons do not necessarily mimic physiological conditions. For example, very high glucose concentrations of up to 25 mM are common in culture media formulations. As TSC is potentially caused by a disruption of the mechanistic target of rapamycin (mTOR) pathway, a main integrator of metabolic information and intracellular signaling, we aimed to examine the impact of different glucose con-centrationsintheculturemediaoncellularphenotypesimplicatedintubercharacteristics.Here,wepresentpre- liminarydatafromapilotstudyexploringcorticalneuronaldifferentiationonhumanembryonicstemcells(hES) harboring a TSC2 knockout mutation (TSC2 − / − ) and an isogenic control line (TSC2+/+). We show that the commonly usedhighglucosemedia profoundlymaskcellular phenotypesinTSC2 − / − culturesduringneuronal differentiation.These phenotypesonly become apparent when differentiating TSC2+/+ and TSC2 − / − cultures in more physiologically relevant conditions of 5 mM glucose suggesting that the careful consideration of culture conditions is vital to ensuring biological relevance and translatability of stem cell models for neurological disor- ders such as TSC. Thisarticleispart oftheSpecial Issue “ Proceedings of the 7th London-InnsbruckColloquium onStatusEp- ilepticus and Acute Seizures".


Introduction
Tuberous sclerosis complex (TSC) is a neurodevelopmental disorder, caused by mutations in the TSC1 or TSC2 genes, that is characterized by tumors in multiple organs [1]. Brain tumors, such as benign cortical tubers, as well as cortical dysorganization often lead to devastating neurological symptoms including autism spectrum disorder, learning disabilities, and seizures [2]. Epilepsy is present in up to 90% of TSC cases [3]. Seizures often start in infancy [4] with multiple seizure types reported and drug resistance in nearly two-thirds of cases [5]. The recent advent of human stem cell-based in vitro models has fueled hope for advances in target discovery and drug developments in TSC. However, stem cell models to study neurological disorders are still in their infancy, thereby necessitating careful consideration of the model characteristics and translational validity. Although stem cell-derived models are now used to study a variety of different brain disorders including TSC [6][7][8], the pitfalls and key characteristics of these models are still to be fully discovered and described. Certain drawbacks, such as a considerable technical variability [9] and functional immaturity of derived neurons [10,11], are already well documented. Furthermore, reliable neuronal differentiation is very dependent on cell culture media, which may support in vitro culture but not necessarily mimic human physiological conditions. Studying epileptogenesis and acute seizures has generally been limited to animal tissue, mostly rodents, through the use of either in vivo models or in vitro preparations. However, research into mechanistic insights of seizure generation can be limited when using rodent models owing to significant differences in neuronal organization and brain development between rodents and humans [12]. Moreover, genetic epilepsy syndromes such as TSC are challenging to study in animal models, since pathogenic mechanisms likely originate from events during early neural development, a phase that differs profoundly between rodents and humans in terms of cell type diversity, proliferation zones, and timescales [13,14]. This translational barrier might be an important reason why mechanisms underlying human epileptogenesis are still not fully understood [15] and may, at least partly, explain why a preventative or disease-modifying antiepileptogenic therapy is not available in clinical practice, despite promising preclinical results [16]. The scientific field is, therefore, increasingly exploring the use of human-based models to better understand molecular, cellular, and developmental principles of epileptogenesis and acute seizure generation.
Stem cells entered research laboratories in the early 1980s with the exploitation of first mouse and, later, human embryonic stem cells (hES) for scientific purposes [17,18]. Since 2006, breakthrough discoveries made by Yamanaka and colleagues enabled the derivation of induced pluripotent stem cells (iPSCs) from adult somatic cells [19] and further differentiation into, theoretically, any human cell type. Thus, neuroscientists now have access to human brain cells from people with epilepsy without being dependent on specimens from brain surgery or autopsies, meaning that human-based in vitro models for acute seizures, epileptogenesis, and chronic epilepsy are potentially accessible. In addition, the advent of precise genome editing tools such as the CRISPR/Cas9 system [20] has made it possible to create human stem cell lines with a specific mutation of interest as well as genetically matched control lines for validating hypotheses relating to causation.
Cortical tubers and the perituberal cortex in TSC have long been implicated in the generation of seizures [21][22][23][24][25]. Cortical tubers are characterized by laminar dysorganization, dysmorphic neurons, eosinophilic giant cells, and severe gliosis [26]. Neither the developmental causes of this cellular dysmorphology nor the functional consequences that potentially underlie the generation of seizure activity are fully understood. Nevertheless, hyperactivation of the mechanistic target of rapamycin (mTOR) pathway is hypothesized to be a major contributor to clinical features in TSC as knockout of the TSC1:TSC2 complex deregulates the mTOR complex 1 (mTORC1) [27], and the mTOR inhibitor everolimus was recently licensed for the treatment of refractory focal epilepsy in the UK [28] and the US.
Neural differentiation protocols follow the timescale of human brain development in vivo so that a stem cell-based model of TSC has the and an isogenic control line (TSC2+/+) were generated at La Roche-Hoffmann Ltd. and NSCs derived from these hES lines were characterized immunologically. No differences could be observed in the expression of the pluripotency markers SOX2, Nanog, and Oct4 in the TSC2+/+ and TSC2−/− hES. Also, TSC2+/+ and TSC2−/− NSCs expressed the NSC markers Nestin and Pax6 in a similar manner. (B) Neuronal differentiation of TSC2+/ + and TSC2−/− hES was performed by dual SMAD inhibition for 80 days. After 10 days, the neuroepithelial sheet consisting of Pax6-positive NSCs had formed in both cultures followed by Pax6-positive neural rosettes around day 18. Rosettes in TSC2−/− cultures tended to be bigger and cell-denser as in wildtype cultures. After 80 days of maturation, both lines generated TUJ1-positive neurons and few GFAP-positive astrocytes. However, TSC2−/− neuronal cultures were found to be denser despite a similar initial plating density. potential to be particularly suited to the study of the neurodevelopmental causes for cellular malfunctioning and glioneuronal imbalance that ultimately might cause seizures in patients with TSC. Here, we present preliminary data to demonstrate that the common culture condition of a very high glucose level in culture media confounds phenotypic findings in a stem cell-based study to model TSC. Our work highlights the importance of fully defined cell culture conditions when studying disease-relevant processes and also suggests an important role for nutrient supply in TSC disease mechanisms.

Human embryonic and neural stem cell lines
We obtained hES harboring a TSC2 deletion (TSC2−/−) and an isogenic control line (TSC2+/+) from La Hoffmann-Roche Limited, which were created by inserting a neomycin selection cassette into either both TSC2 alleles or into the AAVS1 safe harbor locus of the well-characterized hES line SA001, respectively [6]. Human embryonic stem cells were kept on Matrigel-coated (Corning) cell culture plates (Corning) in mTESR (STEMCell Technologies) with daily media changes and regular passaging with Ethylenediamine tetraacetic acid (EDTA) (Invitrogen). We also obtained neural stem cells (NSCs) generated by La Hoffmann-Roche Limited from the TSC2+/+ and TSC2−/− hES described above [6], which were cultured in NSC media on cell culture dishes coated with poly-Lornithine (Sigma-Aldrich) and laminin (Sigma-Aldrich).
No ethical approvals were required for this research, which utilised established stem cell lines. Approval for use of the established embryonic stem cell lines was granted by the Steering Committee of the UK Stem Cell Bank.

Differentiation of TSC2+/+ and TSC2−/− hES
Neuronal differentiation was carried out as first described by Shi et al. [29]. At 100% confluency, hES were induced with neural induction media (NIM) and kept in NIM with daily media changes for 10-12 days. Afterwards, the neuroepithelial sheet was manually passaged in clumps onto cell culture plates coated with 20 μg/ml laminin. Between day 12 and 18, cultures were kept in neural maintenance media (NMM) supplemented with 20 ng/ml bFGF (Peprotech) with media changes every other day. At around day 18, neural rosettes were manually passaged in NMM on 20 μg/ml laminin. Until day 35, full NMM changes were performed every other day, and cells were passaged with accutase (STEMCell Technologies) between day 22 and 25 if they reached 80-100% confluency. At around day 35, a final plating was performed with accutase in NMM on 20 μg/ml laminin at a plating density of 30,000 cells/cm 2 . From then onwards, cells were kept until day 80-100 with full NMM changes every other day.

Growth curves
Growth curves were obtained for TSC2+/+and TSC2−/− NSCs in their high or low glucose culture conditions. Wells were marked with a cross at the bottom, and a picture from each quadrant in the middle of the plate was taken every day between 24 h after plating and reaching 100% confluency. Images were analyzed with ImageJ/FIJI. Doubling time (Dt) was calculated as

Neuronal differentiation protocols
Immunofluorescent characterization of TSC2 +/+ and TSC2 −/− hES and NSC lines showed no obvious morphological differences between the wildtype and the knockout cell lines with regard to pluripotency at the hES stage and NSC markers (Fig. 1A). During neuronal differentiation, both cell lines produced Pax6-positive neuroepithelial cells around day 10 and Pax6-positive radial glial cells at the neural rosette stage around day 18. After 80 days, cortical neurons were generated in both cultures. However, TSC2−/− showed a higher proliferation potential producing denser and bigger rosettes as well as denser neuronal populations (Fig. 1B).

Necessity to use low glucose media to exacerbates mTOR phenotype
Activation of mTORC1 can be measured by the ratio of phosphorylated S6 (pS6) to total S6 ( Fig. 2A). Protein levels for tuberin (gene product of TSC2), pS6, and S6 were detected by western blotting in TSC2+/ + and TSC2 −/− NSCs in either 25 mM or 5 mM glucose media (Fig.  2B). No differences were observed in the pS6/S6 ratio between wildtype and knockout NSCs in their conventional media. However, after 24 h in culture media containing 5 mM glucose, TSC2+/+ NSCs significantly lowered their mTORC1 activation, whereas TSC2−/− NSCs maintained a similar mTOCR1 activity level as in the 25 mM glucose media. As expected, mTORC1 could be inhibited by treatment with 5 nM rapamycin for 24 h in all cultures and media.

Glucose levels affect proliferation and cellular features in TSC2 cultures
We next examined whether cellular phenotypes such as soma size and proliferation rates were also impacted by glucose concentration. Soma size was examined at different timepoints during neuronal cortical differentiation, using media containing either 25 mM or 5 mM glucose, by flow cytometry analysis of the forward scatter (FS) profile (Fig. 3A). TSC2+/+ and TSC2−/− cells showed no differences in cell size during neuronal differentiation in high glucose media. However, reducing the glucose concentration to 5 mM altered cell size at day 18 in both cell lines and at day 80 in TSC2−/− cultures.
We then investigated the impact of glucose on cellular growth of TSC2 +/+ and TSC2 −/− NSCs in either 25 mM or 5 mM glucose media. In low glucose media, TSC2 +/+ NSCs stopped proliferating completely after an initial drop in cell number. In contrast, the growth of TSC2−/− NSCs was not affected by the lower glucose concentration with no change of the Dt after changing to low glucose media (Fig. 3B).

TSC2 knockout affects cellular composition in low glucose media
To detect how TSC2 knockout impacts cell fate choices during neuronal differentiation in the context of different glucose concentrations, we performed cortical neuronal differentiations on TSC2+/+ and TSC2−/ − hES in either 25 mM or 5 mM glucose media and examined the expression of the neuronal marker TUJ1; the glial maker GFAP; the NSC marker Nestin, and the postmitotic neuronal marker HuC/HuD at day 10, day 18, day 50, and day 80. We also characterized cell cycle activity throughout the differentiation by measuring the levels of Ki67 that marks cells actively progressing through the cell cycle.
TSC2 +/+ showed an expected profile for these differentiation markers during neuronal differentiation with a rapid increase in TUJ1 and a decline in Nestin after the neural progenitor state as well as no significant expression of GFAP. Furthermore, 24.4% and 26.2% of the culture were positive for HuC/HuD at day 50 and day 80, respectively. Cell cycle activity peaked after formation of the neural rosettes at the neural progenitor stage and declined as expected at the differentiation stage after day 50. The change in glucose concentration did not alter these marker profiles in TSC2+/+ cultures with the only exception from this being an ongoing cell cycle activity at day 80 in low glucose media (Fig. 5A). In contrast, glucose concentration strongly impacted on cell fate choices in TSC2 −/− cultures that followed the wildtype marker profile in 25 mM glucose media but showed severely disturbed expression of differentiation markers in 5 mM glucose media. The expression of HuC/ HuD was particularly affected and was completely abolished under low glucose conditions at day 80, alongside a marked increase in GFAP expression to 95.8% at the last timepoint (Fig. 4). Moreover, TSC2−/− cells lost cell cycle activity at day 80 in the low glucose media.

Discussion
In this report, we show preliminary data from a stem cell-based approach to the study of TSC highlighting the importance of using culture conditions that are as relevant as possible to the in vivo human physiology. We compared cellular proliferation and cell fate choices during cortical neuronal differentiation of TSC2 +/+ and TSC2 −/− hES in high and low glucose media. Our data indicate that commonly used high glucose concentrations in vitro might mask important cellular phenotypes in a stem cell model of TSC.
In high glucose conditions, we found no phenotypic differences between TSC2+/+ and TSC2−/− NSCs with regard to cell size and differentiation potential. Only the proliferation rate was increased during differentiation in TSC2 knockout cultures. Higher proliferation rates have been reported in TSC2 −/− NSCs previously [6] although other studies could not confirm this observation [30]. However, lowering the glucose concentration to more physiological levels increases the differences in neuronal differentiation potential between TSC2 +/+ and TSC2−/− culturesmore closely resembling the glioneuronal imbalance and disturbed neuronal development that can be found in cortical tubers in vivo.
In patients with TSC, acute seizure activity likely originates from cortical tubers and the perituberal cortex [21][22][23], which are characterized by a glioneuronal imbalance, the presence of dysmorphic neurons, and eosinophilic giant cells [26]. Hyperactivation of the mTOR pathway has been implicated in the development of cortical tubers as mTORC1 has a well-established role in cell proliferation and cell fate choices [31]. However, molecular mechanisms of tuber formation are difficult to study in animal models as, so far, no animal model of TSC develops an equivalent to cortical tubers [13]. Early knockout of either TSC1 or TSC2 in rodents is lethal necessitating a conditional knockout of TSC1 or TSC2 in either neurons or astrocytes at later stages of neurodevelopment in animal models [32]. Furthermore, fundamental differences in cortical development between humans and rodents might also contribute to the absence of cortical tubers in rodent models. Hence, human-based models are likely necessary to study molecular and cellular mechanisms relevant to tuber development and epileptogenesis in patients with TSC.
The differentiation and maintenance of stem cell-based neuronal cultures are now widely adopted in the study of various neurological diseases. However, a thorough characterization of culture conditions and their impact on cell fate choices and cellular function is not well exploredto some extent because media compositions are not often declared and there may be a reluctance to alter systems that appear to enable adequate differentiation. In our model system, we culture The opposite could be observed with regard to GFAP-positivity, which did not differ for TSC2+/+ cultures in both media, whereas GFAP-positivity increased to nearly 95% in TSC2−/− cultures kept at 5 mM glucose. No differences between cell lines and media were found for TUJ1, which was detected in approx. 95% of all cells in all cultures. and differentiate stem cell-derived neurons in fully defined media based on frequently used basal media such as DMEM and Neurobasal media. The basal media contain nonphysiological high glucose concentrations to keep the cultured cells well-supplied with nutrients. However, we have now shown that the high concentration of glucose can mask relevant phenotypes such as cell fate choices, proliferation, and cell size. These phenotypes become apparent when culturing and neuronally differentiating TSC2 −/− hES with a glucose concentration of 5 mM that more closely resembles the glucose levels in the human body.
Further investigations are now warranted to confirm our findings and better understand the mechanisms for disturbed cell fate choices in TSC2 −/− cultures. We might speculate that TSC2 −/− cells have lost their ability to match cell metabolism to nutritional supply by failing to downregulate the mTORC1 pathway. This might then impact on the ability to activate correct signaling pathways in neuronal differentiation. When using conventional media compositions, high glucose in the media may mask this mismatch as TSC2 −/− cells would still have a sufficient supply of glucose for their increased metabolism. As in every model system, it is vital to design the most appropriate model parameters to balance high-quality reproducible results and biological relevance. More broadly, our work highlights the importance of considering and fully reporting protocols and media system ingredients to ensure meaningful and reproducible results [33]. As other limitations of stem cell-based studies (such as immaturity of derived neurons [10], high variability [9], difficulties in obtaining appropriate iPSC control lines, and the challenge to replicate neurodevelopment in 3dimensional structures such as brain organoids [34]) are increasingly addressed, close attention to culture media composition will increasingly support the reliability of such models and improve their future clinical relevance.