Abstract
The RS0 cell line is a rat-derived pheochromocytoma line developed as a model to study pheochromocytoma/paraganglioma caused by hereditary mutations of the SDHB gene. Previous studies demonstrated that xenografts of the RS0 parent tumor replicate characteristics of their human counterparts, including loss of SDHB and upregulation of genes in hypoxia signaling pathways activated by EPAS1/HIF2A. Establishment of the cell line required a low O2 concentration, as cell proliferation was arrested in a traditional cell culture atmosphere of 20% O2. The present study profiled the effects of 20% versus 5% O2 and EPAS1/HIF2A inhibitors on the RS0 cell phenotype and tested how RS0 cells cultured under these influences compare to their parent xenografts and normal rat adrenal medulla. O2 concentration in cell cultures influences almost every aspect of the cells’ biology, most obviously proliferation but also ultrastructure, transcriptome, metabolism and endocrine function. The cells most closely resemble their xenografts when maintained in a low O2 environment, but some differences between the cells in vivo and in vitro are not fully explained. Genes downregulated in high O2 are predominantly associated with the cell cycle, while those upregulated in low O2 include stemness and neuronal progenitor markers that may have contributed to the establishment of the cell line, as well as drug targets expressed in human pheochromocytoma/paraganglioma. Some effects of high O2 are mimicked by EPAS1/HIF2A inhibitors currently considered for treatment of metastatic Sdh-deficient PPGL, while others may be HIF-independent. The cytostatic effect of EPAS1/HIF2A inhibitors is reversible, suggesting possible limits to their usefulness as monotherapies.
Introduction
Pheochromocytomas and paragangliomas (PPGL) are genetically and phenotypically diverse neuroendocrine tumors. At least 40% have a hereditary basis (Fishbein et al. 2017, Flores et al. 2021), and their clinical behavior is strongly determined by the mutated gene. SDHB mutations carry the highest risk of metastasis, estimated at ∼30–40% (Jochmanova et al. 2017, Rijken et al. 2017), which is the most frequent cause of tumor-associated mortality (Dahia et al. 2020). There is no cure for metastatic disease, and the development of treatments is hampered by the lack of experimental models faithfully representing genotype and phenotype (Tischler & Favier 2023). There are no human PPGL cell lines. To address the model gap, in 2020 the Tischler laboratory introduced a rat-derived xenograft and cell line model called RS0 (for Rat SDH-null) developed from a heterozygous rat with a germline deletion in Sdhb (Powers et al. 2020). RS0 cells lost the wild-type Sdhb allele during tumorigenesis and phenotypically resemble succinate dehydrogenase (SDH)-deficient human paragangliomas caused by hereditary SDHB mutations. Fortuitously, a tumor from another rat in the same colony lost the mutant allele by chromothripsis while retaining the wild-type allele, giving rise to an SDH-intact xenograft and cell line called RS1/2 that can serve as a control for RS0 in many respects (Powers et al. 2020).
RS0 and RS1/2 xenografts, which were extensively characterized in our initial paper, provided an in vivo reference point from which to evaluate the relevance of cell culture experiments testing conditions that affect the growth, survival and molecular phenotype of RS0 and RS1/2 cells and their responses to drugs targeting vulnerabilities conferred by SDH deficiency. Both cell lines arose from transplants of their respective primary tumors, similar to human PDX models. Conditions within the xenografts therefore mirrored those under which the primary tumors arose, in contrast to xenografts derived from pre-existing cell lines. An important condition of most solid tissues in vivo is a low oxygen concentration, averaging ∼5% in normal tissues and often dropping to below 1% in corresponding tumors (McKeown 2014). Based on that information, we tested different oxygen concentrations when attempting to establish cell lines from the xenografts and found that RS0 cells would only proliferate in an atmosphere of 5% O2 (Powers et al. 2020). The present studies tested how oxygen influences the proliferation and phenotype of the RS0 and RS1/2 cell lines and how drugs that inhibit hypoxic signaling affect tumor cells maintained in a physiologically relevant oxygen concentration that favors tumor cell growth. The underlying hypothesis is that maintenance of cell cultures in a low oxygen concentration models in vivo tumor growth more faithfully than conventional cultures, allowing for better understanding of tumor biology and facilitating efforts to develop new treatments.
SDH is a dual-function enzyme, serving both as a component of the TCA cycle, where it catalyzes production of fumarate from succinate, and as complex II of the electron transport chain, to which it contributes two electrons from the reaction. SDH-deficient paragangliomas consequently are characterized by altered electron transport and metabolism that includes shunting of TCA cycle intermediates from normal pathways to replenish molecules needed for cell proliferation while also producing enough energy for cell survival. They therefore exist in an abnormal metabolic state that potentially presents selective drug targets (Wang et al. 2022). In addition, the accumulation of succinate inhibits alpha-ketoglutarate-dependent prolyl hydroxylase domain dioxygenases that tag hypoxia-inducible transcription factors for degradation, creating a pseudohypoxic state in which hypoxic signaling pathways are activated abnormally under normoxic conditions. However, it is not clear how SDH-deficient paragangliomas may also be affected by HIF-independent direct effects of oxygen (Chan et al. 2016, Chakraborty et al. 2019) or regulators of HIF activity (Chan et al. 2016). Furthermore, succinate inhibits histone deacetylases and DNA methylases in addition to HIF-prolyl hydroxylases, making it difficult to disentangle the effects of HIFs, oxygen and succinate. Pseudohypoxic signaling pathways mediated by constitutive expression of hypoxia-inducible transcription factor EPAS1/HIF2A have long been considered an attractive drug target, and inhibitors are already being tested in clinical trials (Lussey-Lepoutre et al. 2015, Kamihara et al. 2021). The same inhibitors facilitate distinction between HIF-mediated and HIF-independent oxygen effects.
Materials and methods
Cell cultures
Base medium for both RS0 and RS1/2 cells was RPMI 1640 with GlutaMAX, B27 supplement, basic FGF and EGF (each 20 ng/mL). In addition, RS1/2 but not RS0 medium contains 1% heat-inactivated horse serum and 0.5% fetal bovine serum (Supplementary Table 1 (see section on Supplementary materials given at the end of the article)). Cultures were routinely maintained in Billups-Rothenberg modular incubator chambers with pre-mixed gases consisting of 5% O2 (‘low O2’), 5% CO2 and 90% N2. To test effects of oxygen, cells under these permissive conditions were compared to traditional ‘normoxic’ cultures in 95% air/5% CO2 (∼20% O2, ‘high O2’). The presence or absence of serum in the base medium dates to the time when conditions required to establish each line were determined (Powers et al. 2020). However, after several passages, RS0 cells could tolerate the same low serum concentration, permitting the two lines to be compared under identical conditions to assess confounding effects of serum. All cultures were in a water-saturated atmosphere at 37°C.
Unless otherwise specified, cell proliferation and survival were quantified by direct counting of intact nuclei in a uniform suspension produced by lysing cells in a detergent solution that dissolves cell membranes and cytoplasm (Rukenstein et al. 1991). This method, originally developed to quantify survival of PC12 pheochromocytoma cells (Rukenstein et al. 1991), was employed because it proved to be the most accurate and consistent way to count cells that grow in tight clusters or spheroids. Experiments were typically performed approximately 3 days after trypsinization, allowing time for recovery from trypsin. To assess variability between wells that could have resulted from inadvertent inclusion of incompletely dissociated cell clumps at the onset of experiments, day 0 replicates were plated and then harvested for counting several hours later.
Two inhibitors of EPAS1/HIF2A, PT2399 and PT2977 (also known as belzutifan, MedChem Express, USA), were tested for antiproliferative and cytotoxic effects against RS0 and RS1/2 cells in the presence and absence of serum. Both inhibitors were used at a concentration of 2 uM in view of extensive data documenting on-target effects of PT2399 at that concentration (Cho et al. 2016). In addition, dimethyloxalyl glycine (DMOG, Sigma-Aldrich), an inhibitor of HIF-prolyl hydroxylases (Chan et al. 2016), was utilized at a range of concentrations (10–500 μM) to test whether the inhibitory effect of high O2 is abrogated by pharmacologically induced increases in EPAS1/HIF2A independent of succinate. Both inhibitors were dissolved in DMSO and an equivalent concentration of DMSO was present in controls.
Because there are no human PPGL cell lines, we tested effects of high versus low O2 on primary cultures of these tumors. Three SDH-deficient paragangliomas harboring germline SDHB mutations were dissociated in collagenase as previously reported (Tischler et al. 1984) and plated in 35 mm dishes in a rich medium previously used to optimize survival of primary cultures from an SDH-deficient human gastrointestinal stromal tumor (Powers et al. 2018) (Supplementary Table 1). Heterogeneous cell mixtures containing a total of ∼15,000 cells were plated in replicate 35 mm dishes in an atmosphere containing 1, 5 or 20% O2. For each tumor, one dish was fixed and stained for either tyrosine hydroxylase (TH) or synaptophysin at 4 days and another at 44 days to discriminate the neoplastic cells from other cell types. The 40-day cultures were labeled with bromodeoxyuridine (BrdU) for 3 days before fixation and double-stained for BrdU plus TH or synaptophysin as previously described (Tischler et al. 1992). Tumor cells were quantitated by visual counts of all TH-positive cells under a randomly placed 22 × 22 mm coverslip (Giubellino et al. 2013).
Transcriptome analysis
RNA sequencing was performed by the Tufts Genomics Core Facility. Calculation of RPKM (reads per kilobase of transcript per million mapped reads) and cross-species consensus clustering analysis were as previously described (Powers et al. 2020), with minor modifications. For the present study the most recent release of the Norway Rat genome (mRatBN72) was used for read mapping. This version, 7.2, was released on 11/10/2020 (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_015227675.2/) and is a substantial upgrade over the previous version 6, released July 20, 2014. It holds 41,597 genes compared to 31,595 in version 6 (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_000001895.5/) (Howe et al. 2021), including about 1,000 additional genes with human orthologs. Importantly, for our work, it has much better coverage of several genes highly expressed in RS0 cells and critical to their metabolism but barely detectable in version 6 (e.g., the Eno2 gene encoding ‘neuron-specific enolase’, which is a neuroendocrine marker important for glycolysis and a potential drug target) (Yan et al. 2023). In addition, we now used the STAR aligner rather than Tophat2 for mapping reads. The STAR aligner is faster and gives fewer false positive alignments than Tophat2 (Dobin et al. 2013). For analysis of differential gene expression, genes with less than an average of 50 reads across comparative conditions were removed before calculating log2 values to reduce noise. This left approximately 12,000 mRNAs for analysis. Pathway analysis was performed with Clustergrammer (Fernandez et al. 2017) and EnrichR (Kuleshov et al. 2016), and genes that changed >2Z or <−2Z by comparison to growth in a 5% O2 atmosphere were selected as differentially expressed.
Metabolomics
Metabolomic analysis was performed as previously described (Powers et al. 2018). Briefly, unbiased metabolite profiling was performed using 1H-NMR and 13C-NMR was used to trace the metabolic fates of glucose and the contributions of glycolysis or anaplerotic TCA cycle pathways to the metabolite profile (Bruntz et al. 2017).
Immunoblots
Protein extraction and immunoblotting were performed as previously described (Powers et al. 2000). To allow for probing a wide range of molecular weights, proteins were routinely resolved on 4–15% polyacrylamide gradient gels and blots were sequentially probed, stripped and reprobed with multiple antibodies as indicated in the figure legends.
Statistics
Data are presented as mean ± SEM with individual values displayed on graphs. Statistical analysis was carried out using GraphPad Prism software (GraphPad Software Inc, USA). Significance was tested using an unpaired Student’s t-test, two-tailed distribution, with P < 0.05 set as significant.
Study approval
Patients’ consent was obtained for use of human samples and studies were approved by the Tufts Medical Center Institutional Review Board.
Data availability
RNA-seq data are available through the NCBI Short-Read Archive (SRA) (biproject ID: PRJNA1259954).
Results
Morphological and immunohistochemical characteristics of the RS0 and RS1/2 cell lines
After trypsinization for passaging, RS0 cells initially adhere loosely to a plastic culture dish. As proliferation commences, domes of cells form on top of adherent ones, and after several days all of the cells detach and grow as compact spheroids with no attachment until the next trypsinization. In contrast, RS1/2 cells form aggregates that do not attach or loosely adhere to plastic or collagen (Fig. 1). Both cell lines are uniformly positive for TH. RS0 is negative for Sdhb, while RS1/2 is positive (Fig. 2A). These distinctive growth characteristics and defining phenotype markers have persisted unchanged for well over 40 passages (Fig. 2B) and in all O2 concentrations tested.
Phase contrast photomicrographs typical of RS0 and RS1/2 cell cultures approximately 1 week after trypsinization.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
RS0 and RS1/2 cells maintain distinctive morphology and stably express TH through multiple passages. (A) Immunohistochemical stains on paraffin sections of cultured cells embedded in histogel. (B) Immunoblot of proteins from cells in consecutive passages.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
The most distinctive morphological effects of high versus low O2 are ultrastructural changes in the mitochondria of RS0 cells. At 5% O2, mitochondrial features seen in RS0 xenografts, including increased numbers of irregularly shaped, enlarged or elongated forms with tubular-like or disorganized cristae (Powers et al. 2020), are for the most part retained (Fig. 3A). As in the xenografts, marked mitochondrial swelling and degeneration reported in some other models were not common or consistent findings but were occasionally seen. In contrast, in 20% O2, mitochondria showed extensive severe damage (Fig. 3B). A minor difference from RS0 xenografts was the absence of cytoplasmic vacuoles.
RS0 cells in 5% O2 versus 20% O2. In 5% O2 (A), mitochondria are numerous and enlarged with irregular shapes and disordered cristae (inset at high magnification), but little of the degenerative change that predominates in 20% O2 (B). Bar = 500 nm.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
RS1/2 cells in 5% O2 also retained major features of their corresponding xenografts, including large eccentric norepinephrine-type secretory granules and normal-appearing mitochondria in some cells. However, a broad range of mitochondrial damage was observed in RS1/2 cells even in 5% O2 (Fig. 4).
RS1/2 cells in 5% O2. (A) Portions of adjacent cells illustrating minimally abnormal mitochondria, NE-type secretory granules and focal lipochrome pigment. (B) Different cell in the same culture at 5% O2 with marked mitochondrial swelling and degeneration. Bar = 500 nm.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
Effects of oxygen on cell proliferation and survival
RS0 cells require a low O2 concentration to proliferate. In contrast, RS1/2 cells do proliferate in a traditional cell culture atmosphere of 20% O2 but do so more rapidly at lower O2 concentrations. Figure 5 illustrates differential effects of 5 versus 20% O2 on each cell line in its own optimal medium. The total cytostasis of RS0 cells in 20% O2 persists for at least 21 days and was unchanged in additional experiments when serum was added to the RS0 medium to match the composition of RS1/2 medium (data not shown).
Effect of 5 versus 20% O2 on proliferation of RS0 and RS1/2 cells. Replicate aliquots of cells plated at a density of 10,000 cells/well in 24-well plates on day 0 were maintained in 5% or 20% O2 for 21 days. Quadruplicate wells were counted at weekly intervals. Different scales in the Y-axis reflect different growth rates of the two cell lines. Comparison of means between oxygen groups on individual days: **P < 0.005; ***P < 0.0005.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
Having established RS0 cells as oxygen-averse, we used EPAS1/HIF2A inhibitors to test whether the permissive effect of low O2 on proliferation of these cells is mediated by EPAS1/HIF2A. We first studied the inhibitor PT2399 in view of extensive data documenting on-target effects at the concentration employed (Cho et al. 2016), then similarly tested the newer inhibitor PT2977 (belzutifan). Both inhibitors caused prolonged cytostasis comparable to the effect of 20% O2 (Fig. 6). A small, inconsistent decline in cell number was also seen after 14 days of exposure to belzutifan or 20% O2 in some experiments. For both drugs, cytostasis was reversible and proliferation resumed after washout (Fig. 6). Belzutifan had no effect on survival of RS0 cells in 20% O2 or on proliferation of RS1/2 cells in 5% O2, and the differential effects on RS0 versus RS1/2 cells were unaffected by the presence of 1.5% serum (data not shown).
Growth inhibition by EPAS1 inhibitors is reversible. Triplicate stock dishes of RS0 cells were plated in 5% O2 in the presence or absence of two different EPAS1/HIF2A inhibitors. Cell counts at day 14 were performed to confirm proliferation arrest (initial plating density of the stock dishes indicated by black bar). Replicate aliquots of cells remaining with inhibitor or after inhibitor washout were then plated at 10,000 cells/well in 24-well plates and maintained for up to 14 additional days. Quadruplicate wells were counted on the days shown (technical replicates from two single experiments). Comparison of means between treatment groups on individual days: *P < 0.05; **P < 0.005.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
Given the inhibitory effects of both high O2 and EPAS1/HIF2A inhibitors on cell proliferation, we used the HIF-prolyl hydroxylase inhibitor DMOG to test whether increased HIF signaling alone is sufficient to reverse the inhibitory effect of high O2. DMOG increased EPAS1/HIF2A levels in both RS0 and RS1/2 cells in a dose-dependent manner in high or low O2. Partial reversal of O2-induced cytostasis occurred at the lowest DMOG concentrations eliciting biochemically detectable effects. Higher concentrations continued to increase functional changes, including increased phosphorylation of TH, consistent with the well-known effect of hypoxia on TH activity (Bechmann et al. 2019) (Fig. 7), but with a diminishing effect on proliferation and ultimately toxicity.
DMOG reverses inhibition of RS0 cell proliferation by 20% O2 and mimics effects of low O2 on protein levels. (A) RS0 cells maintained routinely in 5% O2 were transferred to 20% O2 in medium supplemented with the indicated concentrations of DMOG. Cell counts were performed on day 7. (representative experiment showing six technical replicates, experiment repeated three times). Comparison of means between treatment dose and control at 20% O2: *P < 0.05; ***P < 0.0005. (B and C) Immunoblots of protein extract from RS0 cells treated as above and supplemented with DMOG for 24 h. Protein extract from the same cells was run on the two separate blots, which were consecutively stripped and re-probed for the indicated proteins.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
Global effects of oxygen and HIF inhibition
In view of the effects of O2 and HIF inhibition on proliferation of RS0 cells, we tested for parallel effects on transcriptome, protein markers and metabolism to develop an integrated view of global O2 responses and modes of regulation.
Transcriptome
Marked differences were observed between RS0 and R1/2 cells cultured in any O2 concentration, consistent with morphological and functional disparities between the parent xenografts. Our initial paper documenting the pseudohypoxic characteristics of RS0 cells was focused on the xenografts (Powers et al. 2020) and assessed their transcriptomes from the reference point of a table of seven genes derived from an NCBI database comprising the EPAS1/HIF2A regulatory network. These were selected because they showed the greatest differences between RS0 xenograft tissue and normal rat adrenal medulla. The table was revisited here first to assess the effects of the newer rat genome database on our cross-tissue comparisons, then to test RS0 and RS1/2 cells grown in 20 versus 5% O2 to determine which concentration enabled the cultured cells to best replicate the characteristics of corresponding xenografts (Table 1). There were only small absolute differences in reads when analyses were performed with the new rat genome (version 7.2), and ratios of gene expression levels in normal versus xenograft tissues were substantially unchanged. However, more genes had reads in the Rat7.2 genome. Levels at which many individual genes of interest were expressed in vitro most closely corresponded to their in vivo levels when cultures were maintained in 5% rather than 20% O2. However, for some O2-sensitive genes, the high levels of expression in RS0 xenografts were still not reached, suggesting that a still lower O2 concentration might be optimal. In addition, cell culture per se in either high or low O2 increased or decreased the expression of some genes.
Transcriptional effects of oxygen on RSO and RS1/2 cell lines in culture compared to their corresponding xenograft (XG) tissue and rat adrenal medulla (RAM).
| Gene | RAM | RS0-XG | RSO | RS1/2-XG | RS1/2 | ||
|---|---|---|---|---|---|---|---|
| 5% | 20% | 5% | 20% | ||||
| Epas1 | 1 75.42 (80.01) | 890.83 (1,074.63) | 345.83 | 232.46 | 67.70 (76.62) | 2.47 | 0.59 |
| Egln3 | 1.41 (1.44) | 21.12 (23.29) | 45.23 | 2.46 | 0 (0) | 12.71 | 4.21 |
| Bhlhe40 | 10.85 (10.99) | 105.95 (114.0) | 77.05 | 23.32 | 11.96 (11.77) | 14.78 | 13.83 |
| Vegfa | 49.86 (51.43) | 545.02 (607.36) | 288.28 | 84.38 | 60.70 (62.04) | 111.09 | 73.30 |
| Twist1 | 1.82 (1.86) | 13.94 (14.82) | 7.97 | 8.80 | 0.66 (0.50) | 15.22 | 18.27 |
| Serpine1 | 2.58 (2.65) | 232.25 (241.62) | 112.78 | 2.05 | 7.65 (5.9) | 5.01 | 1.90 |
| Adora2a | 2.04 (5.41) | 222.10 (593.88) | 182.38 | 31.92 | 1.21 (2.99) | 2.07 | 0.17 |
Reads per kilobase per million mapped reads (RPKM). Data are based on version 7.2 of the Norway rat genome, except those in parentheses which were previously reported (Powers et al. 2020) based on version 6.
In RS0 cells, 450 transcripts showed decreased expression by Z-score (Cheadle et al. 2003) in the presence of 20% O2, with differences between RS0 and RS1/2 evident at the earliest time point studied (3 days). Three hundred six of these were downregulated in response to both 20% O2 and EPAS1/HIF2A inhibitors (Fig. 8B). These included four of the seven HIF2A target genes in our original reference table (Table 1) (Adora2A, Serpine1, Bhlhe40 and Egln3), while Vegfa was downregulated predominantly by 20% O2 and Twist was essentially unresponsive to high O2 or EPAS1/HIF2A inhibition.
(A) Heat map of the top 100 genes changed up or down by >2Z in cultures of RS0 and RS1/2 cells switched from 5 to 20% O2 for 3 and 14 days (for an interactive version enabling detailed analysis of the indicated genes using multiple platforms see https://maayanlab.cloud/clustergrammer/viz/6765fd2a357ed90010985dca/top*20,100*20combined*20genes*20by*20Z.txt__;JSUlJSU!!CYGGtyGZyw!H0YgGI068guk0idQC5qEVI2qU93LLiXWcBHrTaiOfuPNlbO1U0L0f74n1WMs9-L7WwfzWDe2lu6onop5vaTKr9dG7Scx$). (B) Separate and overlapping responses to 20% O2 or PT2399 by all downregulated genes (Z-score >Z2).
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
A heat map of the top 100 differentially expressed genes in RS0 after a 3-day switch from 5 to 20% O2 compared to the same for RS1/2 or treatment with an EPAS1/HIF2A inhibitor, is shown in Fig. 8A. The effect of holding the cultures to 14 days is to increase levels of differential expression in the RS0 cells. This effect is less apparent for RS1/2 cells. Also notable is that the directional changes in gene expression for RS0 cells either switched to 20% O2 or treated with the inhibitor are largely the same, suggesting that the EPAS1/HIF2A transcription factor accounts for a large portion of the O2 effects.
Genes that changed up or down by more than 2Z were analyzed for roles in reactome pathways (Fabregat et al. 2018). Table 2 shows a heat map for the top ten pathways as scored by P-value for each condition. Not surprisingly, the switch to growth-inhibitory conditions (20% O2 or HIF2i) was dominated by downregulation of genes involved in cell cycle progression. This pertained across both cell lines and conditions. However, there was almost complete discordance between RS0 versus RS1/2 among genes and pathways upregulated after the switch to 20% O2. In RS0 cells, several top-enriched pathways involved neuronal functions such as neuronal development, axon guidance and synaptic transmission, suggesting that the switch to high oxygen induces a partial differentiation response. Gene Ontology analysis leads to the same conclusion (data not shown).
Pathways enriched in 20% O2 and HIF2i-regulated genes in RS0 and RS1/2 cells 1 .
| A. Inhibited genes | |||
|---|---|---|---|
| Pathway | 20% O2 in RS0 2 | HIF2i in RS0 | 20% O2 in RS1/2 |
| Amplification of signal from the kinetochores | 5.26E-24 | 2.69E-39 | 2.31E-17 |
| Amplification of signal from unattached kinetochores via a MAD2 inhibitory signal | 5.26E-24 | 2.69E-39 | 2.31E-17 |
| Cell cycle | 3.96E-60 | 4.49E-87 | 2.77E-27 |
| Cell cycle checkpoints | 5.28E-42 | 6.37E-56 | |
| Cell cycle, mitotic | 4.87E-53 | 7.31E-76 | 3.76E-23 |
| EML4 and NUDC in mitotic spindle formation | 2.01E-22 | 1.80E-40 | 1.17E-17 |
| M phase | 7.79E-23 | 4.67E-38 | 6.30E-18 |
| Mitotic anaphase | 6.82E-16 | ||
| Mitotic metaphase and anaphase | 4.86E-39 | ||
| Mitotic prometaphase | 3.09E-26 | 1.53E-42 | 6.48E-21 |
| Mitotic spindle checkpoint | 5.13E-23 | 6.09E-17 | |
| Resolution of sister chromatid cohesion | 4.52E-26 | 3.43E-44 | 4.26E-19 |
| B. Induced genes | |||
|---|---|---|---|
| Pathway | 20% O2 in RS0 2 | HIF2i in RS0 | 20% O2 in RS1/2 |
| Activation of G protein gated potassium channels | 3.93E-03 | ||
| Axon guidance | 1.14E-04 | 1.02E-03 | |
| FOXO-mediated transcription | 3.45E-03 | ||
| FOXO-mediated transcription of cell death genes | 9.38E-03 | ||
| G protein gated potassium channels | 3.93E-03 | ||
| GAB1 signalosome | 1.06E-02 | ||
| GPCR downstream signaling | 1.87E-03 | 9.02E-04 | |
| Kidney development | 3.58E-04 | ||
| Muscarinic acetylcholine receptors | 1.73E-03 | ||
| Nervous system development | 2.00E-04 | 1.68E-03 | |
| Netrin-1 signaling | 1.39E-03 | ||
| Neuronal system | 2.15E-03 | 4.45E-04 | |
| Neurotransmitter receptors and postsynaptic signal transmission | 4.33E-04 | ||
| PTK6 promotes HIF1A stabilization | 1.73E-03 | ||
| Regulation of commissural axon pathfinding by SLIT and ROBO | 1.51E-04 | 2.63E-04 | |
| Regulation of TBK1, IKKε-mediated activation of IRF3, IRF7 upon TLR3 ligation | 4.43E-03 | ||
| RIP-mediated NFkB activation via ZBP1 | 1.06E-02 | ||
| Signaling by GPCR | 1.34E-03 | ||
| TICAM1-dependent activation of IRF3 IRF7 | 6.21E-03 | ||
| TICAM1, TRAF6-dependent induction of TAK1 complex | 4.43E-03 | ||
| TLR3-mediated TICAM1-dependent programmed cell death | 1.25E-03 | ||
| Toxicity of botulinum toxin type D (botD) | 1.77E-03 | ||
| Toxicity of botulinum toxin type F (botF) | 1.77E-03 | ||
| Transcriptional regulation by the AP-2 (TFAP2) family of transcription factors | 4.44E-03 | ||
| Transmission across chemical synapses | 3.13E-03 | 1.03E-03 | |
Genes that were inhibited (A) or induced (B) by the switch from 5 to 20% O2 or treatment with the HIF2A inhibitor for 3 days were analyzed for enrichment of reactome pathways. Regulated genes analyzed were >2 SD from the mean. The top ten pathways by P value for each condition are shown.
P-values.
Notably, some genes commonly overexpressed in both RS0 cells and SDH-deficient human paragangliomas (Eckardt et al. 2021) (Supplementary Table 2), were increased above in vivo levels by low O2. These include lactate dehydrogenase A (Ldha), increased ∼9-fold over corresponding xenograft tissue in RS0 cells at 5% O2 versus <3-fold at 20% O2. Furthermore, absolute expression levels of Ldha reached in RS1/2 cells, which are not SDH-deficient, were almost the same as in RS0, but the fold increase over xenograft tissue was much higher in RS1/2 cells because of their lower baseline expression. Treatment of cultures maintained at 5% O2 with belzutifan variably mimicked the effect of 20% O2, with some genes affected weakly or not at all. A gene functionally related to Ldha that was both constitutively overexpressed in RS0 xenografts and increased in low O2 was Slc16a1, also known as monocarboxylate transporter 1 (MCT1), a pH-gated outward transporter of lactate and succinate and a potential drug target (Bisbach et al. 2022).
Other important genes selectively affected by O2 concentration and/or EPAS1/HIF2A inhibitors involved endocrine function (Table 3). In both RS0 and RS1/2 cells, these included DOPA decarboxylase, TH, vesicular monoamine transporter 1 (VMAT1) and chromogranin B. Dopamine beta-hydroxylase was almost undetectable in RS0 cells and xenografts, consistent with the predominant production of dopamine in SDH-deficient paragangliomas (Eisenhofer et al. 2012). Choline acetyltransferase (CHAT) expression was essentially undetectable in vivo but appeared at low levels in both RS0 and RS1/2 cell cultures, with minimally higher levels at low O2. However, vesicular acetylcholine transporter (SLC18A3) was not detectable, indicating that if acetylcholine is produced it probably cannot be packaged as a conventional secretory product. In addition, serotonin transporter, which was highly expressed in normal adrenal medulla, was markedly downregulated in culture, while tryptophan hydroxylase 1, which was undetectable in adrenal medulla, was expressed, although at low levels, particularly in RS0 cells in 5% O2 Table 3.
Transcriptional effects (RPKM) of 5 versus 20% oxygen and PT2399 on genes related to endocrine function in RS0 and RS1/2 cells in culture compared to their corresponding xenograft (XG) tissue and rat adrenal medulla (RAM).
| Gene | RAM | RSO-XG | RS1/2-XG | 1 RSO | 1 RS1/2 | 2 RSO (5% O2) | |||
|---|---|---|---|---|---|---|---|---|---|
| 5% | 20% | 5% | 20% | Cntrl | PT2399 | ||||
| Secretory vesicle proteins 3 | |||||||||
| Chga | 10,176.28 | 184.33 | 9,331.32 | 1,143.05 | 993.79 | 4,656.81 | 3,487.53 | 1,104.75 | 1,075.15 |
| Chgb | 6,639.54 | 9,841.81 | 14,979.21 | 5,682.27 | 3,808.20 | 8,715.74 | 6,491.80 | 4,028.57 | 3,622.49 |
| Scg2 | 1,073.87 | 4,730.27 | 842.28 | 1,988.87 | 2,202.30 | 510.84 | 787.93 | 1,329.25 | 1,707.51 |
| Npy | 905.47 | 196.99 | 29.20 | 237.15 | 81.15 | 0.11 | 0 | 46.44 | 54.36 |
| Catecholamine synthesis/metabolism 4 | |||||||||
| Pnmt | 837.63 | 0.34 | 0.97 | 0.19 | 0.13 | 0.08 | 0.06 | 0.10 | 0.12 |
| Th | 2,041.78 | 6,489.11 | 6,975.58 | 2,017.04 | 1,238.07 | 3,791.33 | 4,907.74 | 1,901.90 | 1,053.81 |
| Dbh | 2,218.04 | 7.34 | 8,251.14 | 7.68 | 1.32 | 1,453.02 | 1,357.15 | 8.02 | 6.38 |
| Ddc | 71.05 | 277.22 | 255.13 | 220.80 | 186.53 | 193.67 | 63.30 | 259.65 | 201.45 |
| Maoa | 42.70 | 31.74 | 166.20 | 42.08 | 84.55 | 141.72 | 111.02 | 49.73 | 72.63 |
| Comt | 53.66 | 54.52 | 26.69 | 34.39 | 31.98 | 34.66 | 23.55 | 38.26 | 31.48 |
| Slc18a1 (Vmat1) | 208.23 | 1,131.83 | 24.90 | 302.81 | 180.25 | 83.75 | 23.60 | 456.58 | 145.52 |
| Slc18a2 (Vmat2) | 11.82 | 11.52 | 4.95 | 7.67 | 6.53 | 3.93 | 6.57 | 11.63 | 7.50 |
| Vesicle membrane docking proteins 5 | |||||||||
| Syp | 369.20 | 1,092.80 | 403.64 | 998.60 | 1,051.81 | 433.10 | 480.12 | 1,005.52 | 818.56 |
| Snap25 | 37.91 | 143.39 | 191.64 | 124.01 | 211.34 | 234.89 | 254.00 | 109.79 | 176.81 |
| Vamp1 | 51.94 | 51.86 | 119.80 | 24.10 | 272.52 | 40.94 | 30.20 | 104.01 | 225.32 |
| Acetylcholine and serotonin- related genes 6 | |||||||||
| Chat | 0.41 | 0.34 | 0.28 | 12.24 | 8.64 | 13.75 | 10.14 | 20.50 | 15.91 |
| Slc18a3 (VAChT) | 0.43 | 0 | 0.17 | 0.59 | 0.57 | 0.11 | 0.03 | 0.70 | 0.71 |
| Tph1 | 0 | 2.32 | 0.92 | 2.45 | 0.53 | 0.67 | 1.16 | 0.43 | 0.37 |
| Tph2 | 7.78 | 0 | 1.15 | 0.73 | 0.75 | 1.37 | 0.57 | 0.81 | 1.25 |
| Slc6a4 | 290.51 | 3.95 | 11.89 | 0.06 | 0.30 | 5.11 | 0.89 | 0.14 | 0.28 |
| Htr1a | 0.39 | 0 | 0 | 2.41 | 0.78 | 0.03 | 0.04 | 0 | 0 |
| Htr3a | 2.35 | 2.05 | 1.12 | 8.93 | 19.88 | 5.96 | 1.30 | 12.34 | 18.03 |
RSO and RS1/2 cells maintained in 5% O2 were trypsinized, then plated under the indicated O2 concentration for 14 days.
Effects of PT2399 were tested at 5% O2 in a separate experiment with its own RNA-seq analysis.
Chga, chromogranin A; Chgb, chromogranin B; Scg2, secretogranin II; Npy, neuropeptide Y.
Pnmt, phenylethanolamine-N-methyltransferase; Th, tyrosine hydroxylase; Dbh, dopamine beta-hydroxylase; Ddc, dopa decarboxylase; Maoa, monoamine oxidase A; Comt, catechol-O-methyltransferase; Slc18a1, solute carrier family 18 member A1; Slc18a2, solute carrier family 18 member A2.
Syp, synaptophysin; Snap25, synaptosome associated protein 25; Vamp1, vesicle-associated membrane protein 1.
Chat, choline acetyltransferase; Slc18a3, vesicular acetylcholine transporter; Tph1, tryptophan hydroxylase 1; Tph2, tryptophan hydroxylase 2; Slc6a4, serotonin transporter; Htr1a, 5-hydroxytryptamine receptor 1A; Htr3a, 5-hydroxytryptamine receptor 3A.
To gain insights into mechanisms apparently regulating both cell proliferation and endocrine function, we examined expression of putative stemness and progenitor genes involved in sympathoadrenal development (Kameneva et al. 2021, Bauer & Currie 2024) and expressed in paragangliomas (Oudijk et al. 2015). Genes of interest were identified individually from relevant publications because tissue-specific stemness genes are not represented in reactome pathways or in the GO analysis. In relation to normal adrenal medulla, cultured RS0 and RS1/2 cells showed upregulation or downregulation of several genes in this category, with varied responses to 5 versus 20% O2. These responses were present at 3 days and generally persisted or increased at 14 days (Table 4). However, no specific functions could be ascribed except for the ‘bridge cell’ gene Htr3a, which limits the size of the adrenal medulla during embryogenesis by restricting development of chromaffin cells from neural crest-derived progenitors, maintaining the progenitors in a more proliferative neuroblastic state (Kameneva et al. 2022, Bauer & Currie 2024).
Stemness, progenitor and bridge cell genes in 5 versus 20% O2 in the RS0 and RS1/2 cell lines compared to their corresponding xenograft (XG) tissue and rat adrenal medulla (RAM). Reads per kilobase per million mapped reads (RPKM).
| Gene 1 | RAM | RSO-XG | RS1/2-XG | RSO 2 | RS1/2 2 | RSO (5% O2) 3 | |||
|---|---|---|---|---|---|---|---|---|---|
| 5% | 20% | 5% | 20% | Cntrl | PT2399 | ||||
| Dlk1 | 1,259.49 3 | 241.76 | 1,360.96 | 1,682.92 | 335.67 | 813.12 | 249.44 | 1,210.37 | 458.78 |
| Ngfr | 63.05 | 158.96 | 50.62 | 1.58 | 81.14 | 30.48 | 15.31 | 23.51 | 49.29 |
| Thy1 | 319.97 | 619.84 | 185.17 | 702.17 | 605.74 | 324.89 | 134.37 | 432.64 | 415.87 |
| Ascl1 | 3.61 | 1.27 | 35.65 | 9.78 | 14.78 | 0.08 | 0.63 | 10.91 | 14.40 |
| Htr3a | 2.35 | 2.05 | 1.12 | 8.93 | 19.88 | 5.96 | 1.30 | 12.34 | 18.03 |
| Gata3 | 38.11 | 7.17 | 41.03 | 0.55 | 3.24 | 43.47 | 42.81 | 0.76 | 2.68 |
Dlk1, delta-like non-canonical Notch ligand 1; Ngfr, nerve growth factor receptor; Thy1, Thy-1 cell surface antigen; Ascl1, achaete-scute family bHLH transcription factor 1; Htr3a, 5-hydroxytryptamine receptor 3A; Gata3, GATA binding protein 3.
RSO and RS1/2 cells maintained in 5% O2 were trypsinized, then plated under the indicated O2 concentration for 14 days.
Effects of PT2399 were tested at 5% O2 in a separate experiment with its own RNA-seq analysis.
Metabolism
Culturing RS0 and RS1/2 cells in either high or low O2 led to substantial absolute and relative changes in their overall metabolite profiles, both in comparison to each other (Table 5) and to their corresponding xenografts that we previously reported. In RS0 cells, succinate is the most abundant detectable metabolite in 5% O2 and the second most abundant at 20% O2. This differed from RS0 xenografts, where succinate ranked third after lactate and taurine (Powers et al. 2020). In RS1/2 cells, lactate dropped from being the second most abundant metabolite (after norepinephrine) to 6th in 5% O2 and 12th in 20% O2, while taurine moved near the bottom of the list (Table 5). These changes were attributable to loss of cellular lactate into the culture medium in vitro and absence of mouse-derived taurine in the xenograft model (Powers et al. 2020). Other large changes included a dominant presence of myo-inositol in RS0 cells at 5% and in RS1/2 cells at both 5 and 20% O2 (Table 5), relative to the levels seen in xenografts.
NMR metabolomic profiles of RS0 and RS1/2 cells in 5 versus 20% O2.
| RSO 5% O2 | RSO 20% O2 | RS1/2 5% O2 | RS1/2 20% O2 | ||||
|---|---|---|---|---|---|---|---|
| Succinate | 1 42.03 | Myo-inositol | 49.92 | Myo-inositol | 73.18 | Myo-inositol | 67.78 |
| Glutamate | 23.70 | Succinate | 29.46 | Glutamate | 38.41 | Glutamate | 26.48 |
| Glycine | 14.99 | Glutamate | 26.60 | Glycine | 13.87 | Glycine | 15.33 |
| Myo-inositol | 10.63 | Glycine | 16.78 | IMP | 8.18 | Glutamine | 9.56 |
| Glutamine | 10.33 | Leucine | 14.12 | Glutamine | 7.03 | Oxidized glutathione | 7.86 |
| Proline | 8.91 | Creatine | 5.73 | Lactate | 6.76 | Asparagine | 7.24 |
| Dopamine | 8.00 | Lactate | 5.12 | Oxidized glutathione | 5.57 | Glutathione | 6.73 |
| ATP (and/or ADP) | 7.39 | sn-glycero-3-phosphocholine | 4.50 | AMP | 5.53 | Aspartate | 6.45 |
| Asparagine | 6.58 | Proline | 3.68 | Aspartate | 5.26 | AMP | 6.34 |
| Lactate | 6.58 | O-phosphocholine | 2.66 | Asparagine | 5.22 | sn-glycero-3-phosphocholine | 4.41 |
| Creatine | 5.77 | Acetate | 2.46 | sn-glycero-3-phosphocholine | 5.14 | Proline | 4.36 |
| Leucine | 5.47 | O-acetylcarnitine | 2.05 | Glutathione | 4.98 | Lactate | 4.24 |
| sn-glycero-3-phosphocholine | 4.86 | Betaine | 1.84 | Taurine | 4.15 | Dopamine | 4.13 |
| O-phosphocholine | 4.46 | Inosine | 0.20 | Proline | 3.87 | IMP | 3.56 |
| Oxidized glutathione | 4.05 | Norepinephrine | 3.48 | O-phosphocholine | 3.39 | ||
| IMP | 2.84 | Dopamine | 3.44 | Taurine | 3.00 | ||
| AMP | 2.63 | Creatine | 3.40 | Isoleucine | 2.94 | ||
| Glutathione | 2.63 | O-phosphocholine | 2.92 | Leucine | 2.77 | ||
| Threonine | 2.43 | Alanine | 2.25 | Alanine | 2.66 | ||
| Lysine | 2.13 | Leucine | 1.82 | Creatine | 2.21 | ||
| #26 of 54 succinate | 1.27 | #28 of 54 succinate | 1.41 | ||||
Bold emphasises the level of the substrate for compromised gene in the model cell line.
nmoles/mg of protein.
Tracing the utilization of U-13C glucose revealed that RS1/2 cells readily converted glucose into lactate, alanine (its transaminated metabolite), and products derived from the TCA cycle intermediate α-ketoglutarate, specifically glutamate (Supplementary Fig. 1). Succinate was not detected as a product. The metabolic pathways in RS1/2 cells were not significantly affected by oxygen levels, as the cells cultured in both 20% O2 and 5% O2 showed similar labeling patterns. In contrast, RS0 cells primarily converted glucose to succinate under low O2 conditions, with some conversion to glutamate and lactate, but showed only low levels of succinate labeling in 20% O2.
Glutamate labeling occurs on the C4 carbon through the action of pyruvate dehydrogenase (PDH) and on the C3 carbon through anaplerotic pyruvate carboxylase (PC). In RS1/2 cells, the PDH-to-PC activity ratio remained consistent across both oxygen levels. However, under low O2 conditions, RS0 cells demonstrated a slight increase in PC activity relative to PDH when compared to RS1/2 cells. This shift suggests a greater reliance on anaplerotic flux to replenish TCA cycle intermediates in RS0 cells under low oxygen conditions.
Effects of oxygen on human PPGL cultures
In view of the low O2 requirement for proliferation of RS0 cells, we tested a range of O2 concentrations for effects on growth and survival of dissociated cells from three human paragangliomas caused by hereditary SDHB mutations. There was no proliferation of the neoplastic cells and their populations dwindled over a 40-day period. However, a dose-dependent inverse effect of O2 concentration on survival was observed (Fig. 9).
(A) Effects of low O2 concentrations on survival of dissociated human abdominal paraganglioma cells in primary cultures for 44 days. All three patients had confirmed germline pathogenic SDHB variants. Replicate 35 mm dishes from three tumors maintained at the indicated O2 concentrations were fixed at 4 and 44 days, stained for TH or synaptophysin to identify the neoplastic cells and coverslipped with a 22/22 mm square coverslip. All stained cells under the coverslip were counted. Comparison of means between O2 concentration and 20% O2: *P < 0.05; **P < 0.005. (B) Representative culture showing five neoplastic TH-positive cells (red cytoplasm) in a background of non-neoplastic fibroblasts and endothelial cells at day 44. The culture was labeled with BrdU for 3 days before fixation and showed scattered BrdU-labeled fibroblasts but no labeled TH-positive cells in the entire dish. Bar = 0.5 mm.
Citation: Endocrine-Related Cancer 32, 6; 10.1530/ERC-25-0022
Discussion
The RS0 model was developed for preclinical research aimed to improve the treatment of patients with paragangliomas caused by pathogenic SDHB variants. In that context, in vitro studies of cell lines to identify potential drug targets, test for cytotoxicity and analyze mechanisms of action would be followed by in vivo animal testing in xenografts. Optimally, the molecular and functional characteristics of xenografts would closely resemble those of human tumors to enable selection of the most specific drug targets, and cultured cells would closely resemble corresponding xenografts to minimize confounding effects of the cell culture environment. In addition, drugs should aim for targets more highly expressed in tumor cells than in their normal counterparts to minimize systemic toxicity. Our initial paper (Powers et al. 2020) demonstrated that RS0 xenografts faithfully mirror most of the genetic and phenotypic characteristics of SDH-deficient human paragangliomas. The present studies aimed to define the major influences of cell culture on the RS0 cell line to provide a foundation for further research aligned with the above considerations.
The most salient finding is that O2 concentration in cultures of RS0 cells profoundly influences almost every aspect of their biology, most notably their ability to proliferate, as well as ultrastructure, transcriptome, metabolism and markers of endocrine function. The cultured cells most closely resemble their corresponding xenografts in a low O2 environment, but some differences between the cells in vivo and in vitro are not fully explained. Some effects of high O2 are mimicked by belzutifan or PT2399 and therefore apparently mediated by EPAS1/HIF2A, while others may be HIF-independent (Chakraborty et al. 2019). Further understanding of the roles of HIF signaling may come from studies using HIF-prolyl hydroxylase inhibitors. In this study, partial reversal of O2-induced cytostasis was achieved with DMOG, which is a pan-inhibitor of 2-oxoglutarate-dependent dioxygenases including HIF prolyl hydroxylase. More specific HIF-selective inhibitors of these enzymes, such as roxadustat, may be useful in future investigations. The overall characteristics and O2 responses of RS0 cells are likely to be largely attributable to SDH deficiency because they are, for the most part, different from those of RS1/2 cells, an SDH-intact cell line developed from the same rat colony.
It has been argued for some time that the ∼20% O2 concentration in traditional cell cultures is not physiological and may skew the interpretation of data from cell culture studies (McKeown 2014). Our findings with RS0 cells support that contention and suggest it could have important implications for preclinical drug testing. Testing in a high O2 atmosphere could underestimate or overestimate cytotoxicity depending on the balance between effects of O2 on proliferation and expression of drug targets or other genes. Several druggable targets highly selective for pseudohypoxic paragangliomas and not highly expressed in normal adrenal medulla are downregulated by 20% O2 in RS0 cells. In addition to EPAS1/HIF2A and VEGFA, examples with currently available drugs include ADORA2A (Jing et al. 2023) and LDHA (Maeda et al. 2022). In a contrary example, L1CAM, one of several potential drug targets identified in a recent proteomic screening of human paragangliomas (Vit et al. 2023), is upregulated above in vivo xenograft levels by 20% O2 and is associated with neuronal differentiation. With respect to proliferation, an important finding from our tests of EPAS1/HIF2A inhibitors on RS0 cells is that their effect was predominantly cytostatic and reversible, suggesting they would not permanently cure a tumor if used alone in vivo. Preliminary experiments that we have performed combining HIF inhibitors with several other drugs did not show a benefit of the inhibitors over the other drugs alone (data not shown). However, it will be of interest to continue that line of investigation.
The inhibition of RS0 cell proliferation by a high O2 concentration is not fully explained. However, a low O2 concentration may help to preserve ‘stemness’ (Hammarlund et al. 2018). Stem cells are defined by the ability to both self-renew and differentiate, allowing expansion and diversification of tissues during normal development and maintenance or repair of tissues in adult life. During development, stemness is maintained by tissue hypoxia, while oxic conditions promote differentiation. It has been hypothesized that hypoxia-inducible transcription factors evolved over millennia as a means to maintain stemness in specific anatomic niches and in response to fluctuating oxygen concentrations in an oxic world (Hammarlund et al. 2018). Consistent with that hypothesis, the hypoxic signaling pathways activated by constitutively high HIF2A in SDH-deficient paragangliomas would lead to increased expression of stemness-associated genes (Hammarlund et al. 2018). Putative stemness markers highly expressed in RS0 cells and maintained at high levels in 5% O2 include DLK1/PREF1 and THY1, which are also potential drug targets (Wang et al. 2021, Hamilton et al. 2024) in SDH-deficient human paragangliomas (Oudijk et al. 2015). DLK1/PREF1 is also decreased by HIF2A inhibitors. Our data showing upregulation of stemness genes in both RS0 and RS1/2 cell cultures are consistent with the hypothesis that these genes were required for establishment of the cell lines, but only a few were selectively increased in 5 versus 20% O2 and for some, the reverse was true.
Beyond effects on cell proliferation, our findings show that O2 signaling regulates endocrine phenotype. We previously showed that RS0 xenografts are dopaminergic and RS1/2 noradrenergic, consistent respectively with the profiles of SDH-deficient and SDH-intact human paragangliomas (Powers et al. 2020). Here we show that mRNAs encoding DOPA decarboxylase and dopamine beta-hydroxylase, the immediate biosynthetic enzymes for dopamine and norepinephrine, decrease in cell cultures of RS0 and RS1/2 compared to xenograft tissue but are maintained at higher levels in 5% than in 20% O2. TH mRNA also decreases in cell cultures, but partial reversal of the decrease by low O2 is seen only with RS0. These findings indicate that O2 concentration should be considered when studying mechanisms regulating catecholamine profile.
We also show that O2 concentration and cell culture can affect other aspects of endocrine functions, including qualitative switching from one predominant class of secretory product to another. Questions concerning this type of switch are whether paragangliomas can produce acetylcholine and whether choline acetyltransferase (CHAT) can serve as an immunohistochemical marker, especially for head and neck paragangliomas, which are often SDH-deficient and lacking in TH (Kimura et al. 2021). We have shown that both RS0 and RS1/2 cells can express CHAT mRNA but essentially only in cell culture, where highest levels are associated with lowest O2 (Table 3). Regulated expression of CHAT in cell culture is consistent with the well-known stimulation of acetylcholine synthesis in PC12 pheochromocytoma cells (also rat-derived) by nerve growth factor (Schubert et al. 1977), suggesting that CHAT expression may be associated with acquisition of neuronal traits. Expression of the bridge cell gene Ascl1 in our cell cultures supports this hypothesis, although the somewhat higher expression in 20 versus 5% O2 is not explained. Genes related to serotonin (Table 3) are of similar interest. Serotonin is present in chromaffin cells of the adrenal medulla and extra-adrenal paraganglia (Habeck et al. 1994, Kameneva et al. 2022). In adults, it is derived by uptake via SERT transporters (Habeck et al. 1994) and serves as an inhibitory modulator of catecholamine secretion by acting on 5-HT1A receptors (Bauer & Currie 2024). However, during embryonic development it acts to limit the size of the adrenal medulla by acting on 5-HT3 receptors to restrict the development of chromaffin cells by maintaining their progenitors in a neuroblastic state (Kameneva et al. 2022, Bauer & Currie 2024). 5-HT3 is therefore considered a ‘bridge cell’ marker. Here we showed that serotonin transporter SLC6A4 mRNA is expressed at a moderate level in normal rat adrenal medulla and much lower levels in RS0 and RS1/2 xenografts, becoming almost undetectable in cell culture under most conditions, while the serotonin receptor switches in culture from 5-HT1A to 5-HT3. It is therefore of interest that both 5-HT3 and ASCL1, another bridge cell marker that maintains neural stemness and a proliferative neural phenotype (Parkinson et al. 2022), are upregulated in cell cultures.
Despite similarities of RS0 cells to SDHB-associated human paraganglioma, important differences do exist. A hypoxic mitochondrial phenotype defined by cystic foamy mitochondria devoid or nearly devoid of cristae (Szarek et al. 2015), described in SDH-deficient human paragangliomas and some mouse models (D'Antongiovanni et al. 2017, Douwes Dekker et al. 2003, Rapizzi et al. 2012, Szarek et al. 2015), is not typical of RS0 cells in xenografts or in cell cultures at 5% O2. Rather, RS0 cells resemble immortalized rodent chromaffin cells in which increased mass of mitochondria that are abnormal but retain some function may cause less bioenergetic compromise than would occur with complete loss of mitochondrial function (Kluckova et al. 2020, Al Khazal et al. 2024). This could have implications for selection of antitumoral drugs and targets. However, in view of the cell-to-cell variability of RS0 cells in 5% O2 and acquisition of similar changes by RS1/2 cells in culture but not in vivo, it cannot be ruled out that mitochondrial changes in some studies result from nonphysiological culture conditions. Similar abnormalities could also occur in paragangliomas that lack SDH gene mutations but have acquired decreases in SDH activity (Rapizzi et al. 2012). An additional difference is that human PPGL cells do not proliferate in cell culture, as also shown in other studies (Powers et al. 1998, Bayley et al. 2022). Efficacy of some drugs could therefore be overestimated if tested either against proliferating RS0 cells or against primary human tumor cultures that do not employ a marker to discriminate neoplastic cells from fibroblasts and other proliferating cell types. Our finding that low O2 promotes survival of SDH-deficient human PPGL is novel and requires confirmation in larger series.
In summary, our findings demonstrate that RS0 cells are a unique model for studies of SDHB-associated paraganglioma and more widely for research on O2-responsive traits. They may also provide insights into mechanisms regulating cell lineage determination in the developing sympathetic nervous system (Kameneva et al. 2022, Erickson et al. 2023).
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/ERC-25-0022.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.
Funding
This research was funded by the Paradifference Foundation, the SDHB PheoPara Coalition, the Neuroendocrine Tumor Research Foundation and the PheoPara Alliance.
Author contribution statement
JP and AT conceived the study, designed the experiments and wrote the manuscript. BC and XZ performed RNA sequencing and analyzed the data. JDB performed metabolomic analyses. IL performed electron microscopy and AS-B performed immunohistochemical studies.
Ethics
All tissue samples were collected with informed consent, and the study of the tissue material was approved by the Tufts Medical Center Institutional Review Board.
Clinical collaboration
The authors thank our clinical endocrinology collaborator Dr Ronald Lechan for contributing the human tumor specimens used in this study.
References
Al Khazal FJ , Bhat SM , Zhu Y , et al. 2024 Similar deficiencies, different outcomes: succinate dehydrogenase loss in adrenal medulla vs. fibroblast cell culture models of paraganglioma. Cancer Metab 12 39. (https://doi.org/10.1186/s40170-024-00369-9)
Bauer MB & Currie KPM 2024 Serotonin and the serotonin transporter in the adrenal gland. Vitam Horm 124 39–78. (https://doi.org/10.1016/bs.vh.2023.06.002)
Bayley JP , Rebel HG , Scheurwater K , et al. 2022 Long-term in vitro 2D-culture of SDHB and SDHD-related human paragangliomas and pheochromocytomas. PLoS One 17 e0274478. (https://doi.org/10.1371/journal.pone.0274478)
Bechmann N , Poser I , Seifert V , et al. 2019 Impact of extrinsic and intrinsic hypoxia on catecholamine biosynthesis in absence or presence of Hif2alpha in pheochromocytoma cells. Cancers 11 594. (https://doi.org/10.3390/cancers11050594)
Bisbach CM , Hass DT , Thomas ED , et al. 2022 Monocarboxylate transporter 1 (MCT1) mediates succinate export in the retina. Invest Ophthalmol Vis Sci 63 1. (https://doi.org/10.1167/iovs.63.4.1)
Bruntz RC , Lane AN , Higashi RM , et al. 2017 Exploring cancer metabolism using stable isotope-resolved metabolomics (SIRM). J Biol Chem 292 11601–11609.
Chakraborty AA , Laukka T , Myllykoski M , et al. 2019 Histone demethylase KDM6A directly senses oxygen to control chromatin and cell fate. Science 363 1217–1222. (https://doi.org/10.1126/science.aaw1026)
Chan MC , Ilott NE , Schödel J , et al. 2016 Tuning the transcriptional response to hypoxia by inhibiting hypoxia-inducible factor (HIF) prolyl and asparaginyl hydroxylases. J Biol Chem 291 20661–20673. (https://doi.org/10.1074/jbc.m116.749291)
Cheadle C , Vawter MP , Freed WJ , et al. 2003 Analysis of microarray data using Z score transformation. J Mol Diagn 5 73–81. (https://doi.org/10.1016/s1525-1578(10)60455-2)
Cho H , Du X , Rizzi JP , et al. 2016 On-target efficacy of a HIF-2alpha antagonist in preclinical kidney cancer models. Nature 539 107–111. (https://doi.org/10.1038/nature19795)
D'Antongiovanni V , Martinelli S , Richter S , et al. 2017 The microenvironment induces collective migration in SDHB-silenced mouse pheochromocytoma spheroids. Endocr Relat Cancer 24 555–564. (https://doi.org/10.1530/erc-17-0212)
Dahia PLM , Clifton-Bligh R , Gimenez-Roqueplo AP , et al. 2020 Hereditary endocrine tumours: current state-of-the-art and research opportunities: metastatic pheochromocytomas and paragangliomas: proceedings of the MEN2019 workshop. Endocr Relat Cancer 27 T41–t52. (https://doi.org/10.1530/erc-19-0435)
Dobin A , Davis CA , Schlesinger F , et al. 2013 STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29 15–21. (https://doi.org/10.1093/bioinformatics/bts635)
Douwes Dekker PB , Hogendoorn PC , Kuipers-Dijkshoorn N , et al. 2003 SDHD mutations in head and neck paragangliomas result in destabilization of complex II in the mitochondrial respiratory chain with loss of enzymatic activity and abnormal mitochondrial morphology. J Pathol 201 480–486. (https://doi.org/10.1002/path.1461)
Eckardt L , Prange-Barczynska M , Hodson EJ , et al. 2021 Developmental role of PHD2 in the pathogenesis of pseudohypoxic pheochromocytoma. Endocr Relat Cancer 28 757–772. (https://doi.org/10.1530/erc-21-0211)
Eisenhofer G , Lenders JW , Siegert G , et al. 2012 Plasma methoxytyramine: a novel biomarker of metastatic pheochromocytoma and paraganglioma in relation to established risk factors of tumour size, location and SDHB mutation status. Eur J Cancer 48 1739–1749. (https://doi.org/10.1016/j.ejca.2011.07.016)
Erickson AG , Kameneva P & Adameyko I 2023 The transcriptional portraits of the neural crest at the individual cell level. Semin Cell Dev Biol 138 68–80. (https://doi.org/10.1016/j.semcdb.2022.02.017)
Fabregat A , Jupe S , Matthews L , et al. 2018 The reactome pathway knowledgebase. Nucleic Acids Res 46 D649–D655. (https://doi.org/10.1093/nar/gkx1132)
Fernandez NF , Gundersen GW , Rahman A , et al. 2017 Clustergrammer, a web-based heatmap visualization and analysis tool for high-dimensional biological data. Sci Data 4 170151. (https://doi.org/10.1038/sdata.2017.151)
Fishbein L , Leshchiner I , Walter V , et al. 2017 Comprehensive molecular characterization of pheochromocytoma and paraganglioma. Cancer Cell 31 181–193. (https://doi.org/10.1016/j.ccell.2017.01.001)
Flores SK , Estrada-Zuniga CM , Thallapureddy K , et al. 2021 Insights into mechanisms of pheochromocytomas and paragangliomas driven by known or new genetic drivers. Cancers 13 4602. (https://doi.org/10.3390/cancers13184602)
Giubellino A , Bullova P , Nolting S , et al. 2013 Combined inhibition of mTORC1 and mTORC2 signaling pathways is a promising therapeutic option in inhibiting pheochromocytoma tumor growth: in vitro and in vivo studies in female athymic nude mice. Endocrinology 154 646–655. (https://doi.org/10.1210/en.2012-1854)
Habeck JO , Pallot DJ & Kummer W 1994 Serotonin immunoreactivity in the carotid body of adult humans. Histol Histopathol 9 227–232.
Hamilton AK , Radaoui AB , Tsang M , et al. 2024 A proteogenomic surfaceome study identifies DLK1 as an immunotherapeutic target in neuroblastoma. Cancer Cell 42 1970–1982.e7. (https://doi.org/10.1016/j.ccell.2024.10.003)
Hammarlund EU , von Stedingk K & Pahlman S 2018 Refined control of cell stemness allowed animal evolution in the oxic realm. Nat Ecol Evol 2 220–228. (https://doi.org/10.1038/s41559-017-0410-5)
Howe K , Dwinell M , Shimoyama M , et al. 2021 The genome sequence of the Norway rat, Rattus norvegicus Berkenhout 1769. Wellcome Open Res 6 118. (https://doi.org/10.12688/wellcomeopenres.16854.1)
Jing N , Zhang K , Chen X , et al. 2023 ADORA2A-driven proline synthesis triggers epigenetic reprogramming in neuroendocrine prostate and lung cancers. J Clin Invest 133 e168670. (https://doi.org/10.1172/jci168670)
Jochmanova I , Wolf KI , King KS , et al. 2017 SDHB-related pheochromocytoma and paraganglioma penetrance and genotype-phenotype correlations. J Cancer Res Clin Oncol 143 1421–1435. (https://doi.org/10.1007/s00432-017-2397-3)
Kameneva P , Artemov AV , Kastriti ME , et al. 2021 Single-cell transcriptomics of human embryos identifies multiple sympathoblast lineages with potential implications for neuroblastoma origin. Nat Genet 53 694–706. (https://doi.org/10.1038/s41588-021-00818-x)
Kameneva P , Melnikova VI , Kastriti ME , et al. 2022 Serotonin limits generation of chromaffin cells during adrenal organ development. Nat Commun 13 2901. (https://doi.org/10.1038/s41467-022-30438-w)
Kamihara J , Hamilton KV , Pollard JA , et al. 2021 Belzutifan, a potent HIF2alpha inhibitor, in the Pacak-Zhuang syndrome. N Engl J Med 385 2059–2065. (https://doi.org/10.1056/nejmoa2110051)
Kimura N , Shiga K , Kaneko KI , et al. 2021 Immunohistochemical expression of choline acetyltransferase and catecholamine-synthesizing enzymes in head-and-neck and thoracoabdominal paragangliomas and pheochromocytomas. Endocr Pathol 32 442–451. (https://doi.org/10.1007/s12022-021-09694-x)
Kluckova K , Thakker A , Vettore L , et al. 2020 Succinate dehydrogenase deficiency in a chromaffin cell model retains metabolic fitness through the maintenance of mitochondrial NADH oxidoreductase function. FASEB J 34 303–315. (https://doi.org/10.1096/fj.201901456r)
Kuleshov MV , Jones MR , Rouillard AD , et al. 2016 Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res 44 W90–W97. (https://doi.org/10.1093/nar/gkw377)
Lussey-Lepoutre C , Hollinshead KE , Ludwig C , et al. 2015 Loss of succinate dehydrogenase activity results in dependency on pyruvate carboxylation for cellular anabolism. Nat Commun 6 8784. (https://doi.org/10.1038/ncomms9784)
Maeda M , Ko M , Mane MM , et al. 2022 Genetic and drug inhibition of LDH-A: effects on murine gliomas. Cancers 14 2306. (https://doi.org/10.3390/cancers14092306)
McKeown SR 2014 Defining normoxia, physoxia and hypoxia in tumours-implications for treatment response. Br J Radiol 87 20130676. (https://doi.org/10.1259/bjr.20130676)
Oudijk L , Neuhofer CM , Lichtenauer UD , et al. 2015 Immunohistochemical expression of stem cell markers in pheochromocytomas/paragangliomas is associated with SDHx mutations. Eur J Endocrinol 173 43–52. (https://doi.org/10.1530/eje-14-1164)
Parkinson LM , Gillen SL , Woods LM , et al. 2022 The proneural transcription factor ASCL1 regulates cell proliferation and primes for differentiation in neuroblastoma. Front Cell Dev Biol 10 942579. (https://doi.org/10.3389/fcell.2022.942579)
Powers JF , Tsokas P & Tischler AS 1998 The ret-activating ligand GDNF is differentiative and not mitogenic for normal and neoplastic human chromaffin cells in vitro. Endocr Pathol 9 325–331. (https://doi.org/10.1007/bf02739692)
Powers JF , Evinger MJ , Tsokas P , et al. 2000 Pheochromocytoma cell lines from heterozygous neurofibromatosis knockout mice. Cell Tissue Res 302 309–320. (https://doi.org/10.1007/s004410000290)
Powers JF , Cochran B , Baleja JD , et al. 2018 A unique model for SDH-deficient GIST: an endocrine-related cancer. Endocr Relat Cancer 25 943–954. (https://doi.org/10.1530/erc-18-0115)
Powers JF , Cochran B , Baleja JD , et al. 2020 A xenograft and cell line model of SDH-deficient pheochromocytoma derived from Sdhb+/− rats. Endocr Relat Cancer 27 337–354. (https://doi.org/10.1530/erc-19-0474e)
Rapizzi E , Ercolino T , Canu L , et al. 2012 Mitochondrial function and content in pheochromocytoma/paraganglioma of succinate dehydrogenase mutation carriers. Endocr Relat Cancer 19 261–269. (https://doi.org/10.1530/erc-11-0263)
Rijken JA , Niemeijer ND , Jonker MA , et al. 2017 The penetrance of paraganglioma and pheochromocytoma in SDHB germline mutation carriers. Clin Genet 93 60–66. (https://doi.org/10.1111/cge.13055)
Rukenstein A , Rydel RE & Greene LA 1991 Multiple agents rescue PC12 cells from serum-free cell death by translation- and transcription-independent mechanisms. J Neurosci 11 2552–2563. (https://doi.org/10.1523/jneurosci.11-08-02552.1991)
Schubert D , Heinemann S & Kidokoro Y 1977 Cholinergic metabolism and synapse formation by a rat nerve cell line. Proc Natl Acad Sci U S A 74 2579–2583. (https://doi.org/10.1073/pnas.74.6.2579)
Szarek E , Ball ER , Imperiale A , et al. 2015 Carney triad, SDH-deficient tumors, and Sdhb+/− mice share abnormal mitochondria. Endocr Relat Cancer 22 345–352. (https://doi.org/10.1530/erc-15-0069)
Tischler AS & Favier J 2023 Progress and challenges in experimental models for pheochromocytoma and paraganglioma. Endocr Relat Cancer 30 e220405. (https://doi.org/10.1530/erc-22-0405)
Tischler AS , Lee YC , Perlman RL , et al. 1984 Production of “ectopic” vasoactive intestinal peptide-like and neurotensin-like immunoreactivity in human pheochromocytoma cell cultures. J Neurosci 4 1398–1404. (https://doi.org/10.1523/jneurosci.04-05-01398.1984)
Tischler AS , Ruzicka LA & Riseberg JC 1992 Immunocytochemical analysis of chromaffin cell proliferation in vitro. J Histochem Cytochem 40 1043–1045. (https://doi.org/10.1177/40.7.1351491)
Vit O , Talacko P , Musil Z , et al. 2023 Identification of potential molecular targets for the treatment of cluster 1 human pheochromocytoma and paraganglioma via comprehensive proteomic characterization. Clin Proteomics 20 39. (https://doi.org/10.1186/s12014-023-09428-7)
Wang H , Li X , Lai LA , et al. 2021 X-aptamers targeting Thy-1 membrane glycoprotein in pancreatic ductal adenocarcinoma. Biochimie 181 25–33. (https://doi.org/10.1016/j.biochi.2020.11.018)
Wang K , Crona J , Beuschlein F , et al. 2022 Targeted therapies in pheochromocytoma and paraganglioma. J Clin Endocrinol Metab 107 2963–2972. (https://doi.org/10.1210/clinem/dgac471)
Yan VC , Barekatain Y , Lin YH , et al. 2023 Comparative pharmacology of a bis-pivaloyloxymethyl phosphonate prodrug inhibitor of enolase after oral and parenteral administration. ACS Pharmacol Transl Sci 6 245–252. (https://doi.org/10.1021/acsptsci.2c00216)
