SDHD: A Mitochondrial Respiratory Chain Defect that Causes PGL
Recently, PGL1 was identified to be the SDHD gene by using positional cloning methods.40 The SDHD gene encodes the small subunit D of cytochrome b558 of the mitochondrial respiratory chain complex II (succinate: ubiquinone oxidoreductase). SDHD was found to map in the chromosomal segment defined by haplotype and recombinant analyses in families and was an immediate candidate, given the role of the carotid body as a sensory organ for arterial oxygen levels. The 103 amino acids of the mature SDHD-protein (without the leader signal-peptide) are encoded by 4 exons, which cover about 10 kb genomic DNA.41 Germ-line mutations were found in all families that were previously linked to PGL1. The mutations found to date include nonsense and frameshifting mutations, some occurring in the mitochondrial signal peptide, or missense mutations replacing evolutionary conserved amino acids (Fig. 40.1-3). In the Dutch population, in which a strong founder effect had already been noted, over 90 percent of all hereditary cases were found to be caused by either of two missense mutations.42 Germ-line SDHD mutations have also been detected in sporadic and familial nonsyndromic pheochromocytomas.40,43
Gene organization of SDHD and position of disease-related mutations in HN-paraganglioma families. ATG and TGA represent the start and stop codons for protein-translation, respectively. Exons 1–4 are 62, 117, 145, and 989 bp, respectively. Mutation data compiled from refs 40, 42, 43, and 58–60.
In four HN-paraganglioma patients with a germ-line mutation, DNA from flow-sorted diploid and aneuploid tumor cell fractions could be examined. In the aneuploid fraction, only the mutation could be demonstrated to be present. The maternally derived wild-type allele was lost. These results are in agreement with Knudson's two-hit hypothesis for the inactivation of a tumor-suppressor gene. Accordingly, SDHD is the first tumor suppressor gene to exert its normal cellular function in the mitochondria. Knudson's hypothesis also predicts that acquired (or “somatic”) mutations would occur in SDHD, in truly sporadic HN-paragangliomas. One such mutation has been reported in 42 analyzed pheochromocytomas, 40,43 but not yet in HN-paragangliomas.
Other Complex II Proteins
The mitochondrial complex II is composed of four peptides: the flavoprotein and iron-sulfur protein of the succinate dehydrogenase (SDHA and SDHB, respectively), and two integral membrane proteins (SDHC and SDHD) that anchor the enzymatic subunits to the inner-mitochondrial membrane. The link between complex II dysfunction and tumorigenesis was further supported by two findings. First, the HN-paragangliomas in the German family linked to PGL3 were found to be due to a mutation in SDHC.44 This mutation destroys the start-codon ATG of the gene, which leads to a rapid degradation of the mutant transcript. Second, SDHB mutations were detected in two of five kindreds with familial pheochromocytoma, in two of three kindreds with pheochromocytoma and paraganglioma, and in one of 24 sporadic pheochromocytomas.45 Further work is now required to establish the proportions of HN-paraganglioma families carrying mutations in SDHB, SDHC, or SDHD.
At the moment, it is not clear how a defect in a protein that is part of complex II of the respiratory chain can cause tumor growth specifically in HN-paraganglia. It is possible that complex II dysfunction in paraganglia cells interferes with their physiological function as chemoreceptors for blood oxygen pressure. In rats, the rise in partial pressure of oxygen in arterial blood around the time of birth is critical for normal maturation and function of the type I carotid body cells. This is accompanied by a concomitant rise in hypoxia sensitivity (reviewed in Ref.3). Chronic hypoxia after birth (imposed by maintaining the rats in a low-oxygen environment) inhibits this maturation and results in glomus cell hypertrophy. On this basis, it is tempting to speculate that in humans, the loss of SDHD mimics the chronic hypoxic state, despite the actual normoxia conditions in the blood. As a result, the carotid body paraganglia become persistently hyperplastic, 2,46 which increases the chance of acquiring further gene mutations and eventual genuine monoclonal tumor growth (Fig. 40.1-4). Hyperplasia of glomus bodies has been noted in patients with chronic hypoxemia.47 Increased hyperplasia, as well as increased incidence of HN-paragangliomas, has been noted in individuals living for prolonged time at very high altitudes in the Andes of Peru.48 Finally, two-thirds of HN-paragangliomas express vascular endothelial growth factor (VEGF), specifically in the chief cells.49 VEGF is induced by hypoxia, and its up-regulation in HN-paragangliomas may explain their hypervascularity.
Model for the role of the mitochondrial complex II in the development of a carotid body tumor. Normal maturation of the type I carotid body cells, accompanied by a rise in hypoxia sensitivity, occurs simultaneously with the rise in partial pressure of oxygen in arterial blood around the time of birth. Failed maturation (caused by chronic hypoxia or a gene-defect in SDHB, -C, or -D) results in glomus cell hypertrophy. A persistent hyperplastic state increases the chance of acquiring further gene mutations and monoclonal tumor growth (in dark grey).
The exciting corollary of this hypothesis is that mitochondrial complex II is involved in oxygen sensing, for which some experimental evidence is in fact beginning to emerge (reviewed in Ref.50). In response to hypoxia, mitochondria increase their levels of reactive oxygen species, which apparently act as second messengers in a variety of cell types. Moreover, some of the signaling cascades thus elicited appear to participate in the control of cell death pathways (apoptosis). It is thus possible that the carotid body hypertrophy seen in individuals living at high altitudes, and the slow developmental pattern of carotid body tumors in SDHD carriers, is caused by a defect in this apoptotic pathway. Indeed, one study observed that the chief cells of most carotid body tumors express moderate to high levels of the antiapoptotic protein bcl-2. 51 Yet this hypothesis doesn't explain tumorigenesis in the paraganglia other than those at the carotid bifurcation, such as those in the jugular bulb and middle ear. It also remains to be seen whether noninherited sporadic HN-paragangliomas arise by a maturation defect in hypoxia sensing.
SDHD and the Inheritance Pattern of HN-Paragangliomas
The simplest model explaining the inheritance pattern of PGL1-linked HN-paragangliomas is one in which the maternally derived allele is not expressed. After the molecular nature of the gene-defect was discovered, this model predicted that SDHD would behave as a maternally imprinted gene. However, measurements of allele-specific mRNA expression patterns in fetal brain and kidney, and adult brain and blood lymphocytes, show that both alleles are expressed at equal levels.40 The possibility remains that the allele-specific expression of SDHD is specific for the paraganglia cells in the head and neck region. Investigating this is hampered by the size of normal human paraganglia and the difficulty in obtaining these tissues postmortem. Nonetheless, several arguments speak against a classic genomic imprinting model for SDHD. First, the preferential loss of the normal maternal SDHD allele in tumors suggests that it is expressed in paraganglia cells and that its loss provides a selective growth advantage in the development of paraganglioma. By analogy, loss of the wild-type allele was also observed in two tumors caused by SDHC, and, indeed, this family showed no evidence for maternal imprinting.44 Second, if all individuals would inherit an imprinted inactive maternal copy from their mother, Knudson's two-hit hypothesis would effectively be reduced to a one-hit variant, because only a single mutation event would be needed to inactivate the paternal allele and to initiate tumor growth. This is at odds with the rarity of HN-paragangliomas in the general population. Third, SDHD does not map to a chromosomal region known to harbor imprinted genes, or syntenic to mouse chromosome regions that are known to contain genes with parent-of-origin effects. It has been suggested that the SDHD gene is not completely imprinted in normal paraganglia or that secondary relaxation of imprinting occurs prior to tumor formation.40 Even more complex models, invoking a linked, oppositely imprinted gene whose inactivation determines tumor progression, can be conjured up.
Can anything be gleaned from the fact that mitochondria play an essential role as energy suppliers for the cell, while having an exclusive maternal inheritance? Here, the available information is also puzzling. Genetic defects are now known in all subunits of complex II (Table 40.1-1). Mutations in SDHA are recessive, but are found in a small proportion of patients affected with the Leigh syndrome, 52,53 which is characterized by the presence of developmental delay and lactic acidosis and by a mean life expectancy of less than 5 years (MIM 256000). By contrast, mutations in SDHB, SDHC and SDHD are dominant, causing tumor susceptibility in carriers, although the imprinting phenomenon is seen only for SDHD. Knowledge on the effect of these mutations on mitochondrial function is essentially still lacking, but all can apparently be maternally transmitted without reduction of fitness. If all these mutations simply “inactivate” complex II, then one would predict some Leigh syndrome patients to be compound heterozygote or homozygote for SDHB, SDHC or SDHD mutations. Conversely, heterozygote carriers of an SDHA mutation would be at risk for developing HN-paragangliomas. Neither has been found to date, suggesting that SDHA, -B, -C, and -D mutations have differential effects on mitochondrial function. Although we know little about the oxidative requirements of primary and haploid secondary oocytes, it is conceivable that those carrying an SDHD mutation are specifically capable of correcting the defect by an epigenetic mechanism affecting the expression of one or more other genes. This expression pattern might persist throughout adult life, effectively suppressing the development of paraganglioma.
Table 40.1-1: Inherited Defects in Mitochondrial Complex II Subunits |Favorite Table|Download (.pdf) Table 40.1-1: Inherited Defects in Mitochondrial Complex II Subunits
|Subunit (chrom. location) ||Syndrome ||Inheritance pattern ||Heterozygous carriers |
|SDHA (5p15) ||Leigh syndrome ||Recessive ||Unaffected |
|SDHB (1p36) ||Familial pheochromocytoma with or without paraganglioma ||Dominant ||Susceptible to PHEO (and PGL) |
|SDHC (1q21) ||HN-paragangliomas ||Dominant ||Susceptible to PGL |
|SDHD (11q22.3) ||HN-paragangliomas ||Dominant with imprinting ||Susceptible to PGL (and PHEO) |
Specific Genes Involved in Sporadic HN-Paragangliomas.
As explained, the definition of sporadic HN-paragangliomas is somewhat problematic given their benign clinical behavior and typical inheritance pattern. In the Dutch population, inherited cases appear to comprise a much higher proportion than previously estimated.8 Accordingly, the two SDHD founder mutations were found in the germ line of 20 of 55 (36 percent) Dutch patients who had previously been classified as “sporadic” on the basis of lack of an apparent family history.42 Ten of the isolated patients had multiple paragangliomas, and in eight of these patients, SDHD germ-line mutations were found, indicating that multicentricity is a strong predictive factor for the hereditary nature of the disorder in isolated patients. Further work on SDHD mutation prevalence in patients from other ethnicities should discover whether or not such high proportions are a peculiarity of the founder effect in the Dutch population.
Regardless of the genetic origin of HN-paraganglioma, other genes are likely involved in the progression toward clinically manifest disease. The glomus jugulare tumor of one of two affected brothers was found to contain missense mutations in TP53 and CDKN2A, 54 but these may have been induced by radiotherapy. The clinical association between pheochromocytoma and HN-paragangliomas, although weak, might point to common genetic etiology. Pheochromocytomas occur in the context of a number of cancer syndromes, including multiple endocrine neoplasia (MEN1 and MEN2), von Hippel-Lindau disease (VHL), and neurofibromatosis type 1 (NF1). Genes underlying these syndromes are known, and thus form candidates for their involvement in HN-paragangliomas. The RET oncogene, causing MEN2, does not seem to play a role in HN-paraganglioma.55 Another study found LOH at 3p21 in two of four informative cases, but not including the VHL locus at 3p25.56