TY - JOUR
AU - Feustel, Paul J.
AB - After Friedreich’s description in 1877, depletion of myelinated fibers in the dorsal columns, dorsal spinocerebellar and lateral corticospinal tracts, and neuronal loss in the dorsal nuclei of Clarke columns were considered unique and essential neuropathological features of Friedreich ataxia (FA). Lack of large neurons in dorsal root ganglia (DRG), thinning of dorsal roots (DR), and poor myelination in sensory nerves are now recognized as key components of FA. Here, we measured cross-sectional areas of the mid-thoracic spinal cord (SC) and neuronal sizes in lumbosacral DRG of 24 genetically confirmed FA cases. Mean thoracic SC areas in FA (24.17 mm2) were significantly smaller than those in 12 normal controls (37.5 mm2); DRG neuron perikarya in FA (1362 µm2) were also significantly smaller than normal (2004 µm2). DRG neuron sizes were not correlated with SC areas. The FA patients included a wide range of disease onset and duration suggesting that the SC undergoes growth arrest early and remains abnormally small throughout life. Immunohistochemistry for phosphorylated neurofilament protein, peripheral myelin protein 22, and myelin proteolipid protein confirmed chaotic transition of axons into the SC in DR entry zones. We conclude that smaller SC areas and lack of large DRG neurons indicate hypoplasia rather than atrophy in FA. Atrophy, Dorsal roots, Dorsal root ganglion, Friedreich ataxia, Hypoplasia, Spinal cord. INTRODUCTION Since Friedreich’s illustration in 1877 of dorsal column degeneration in the disease that now bears his name (1), many generations of medical trainees have accepted that the most important abnormalities reside in the spinal cord. Beyond the dorsal columns, the disease also affects the dorsal spinocerebellar and lateral corticospinal tracts, and the dorsal nuclei of Clarke columns. While Friedreich was aware of thinning of the dorsal spinal roots, he wrote that dorsal root ganglia (DRG) were normal (1). Remarkably, he observed that the nerve fibers in the dorsal roots (DR) were more abundant than normal and rather thin, displaying only rare thicker fibers similar to those in the intact ventral roots. Over the ensuing years, it became clear that DRG are a critical target of the disease, and that fiber loss in the dorsal columns and neuronal atrophy of the dorsal nuclei could be secondary. The neuropathological phenotype also includes hypoplasia or atrophy of Betz cells of the motor cortex that matches the invariable degeneration of the lateral corticospinal tracts. Although the normal dentate nucleus (DN) receives sparse collaterals from mossy fibers of spinal origin, the severe lesion of the DN in Friedreich ataxia (FA) is not obviously related to the damaged spinal cord. Irrespective of disease onset, the spinal cord in FA is thinner than normal (2), and the bony spinal canal in FA patients is abnormally narrow on plain radiographs (3). In retrospect, this old observation gains significance because of more recent measurements of magnetic resonance images (4–6). The narrow spinal canal on plain X-ray films in FA (3) must be interpreted as the result of incomplete anteroposterior and transverse enlargement of the spinal cord during growth. Therefore, thinning of the spinal cord in FA may be more appropriately termed hypoplasia, although it is likely that inflammation and invasion of DRG neurons continue to damage afferent fibers in DR (7). It is of interest that Friedreich had already proposed in 1877 that smallness of the clavae (gracile and cuneate nuclei) was developmental (1). The neurons of DRG, satellite cells, and Schwann cells of DR and peripheral nerves derive from the neural crest (8), although boundary cap cells may also contribute to the constituents of DRG and proximal and distal processes (9). Axons deriving from DRG neurons grow into the periphery where they become sensory nerves, into the gray matter of the spinal cord, and, over remarkably long distances, into the nuclei of the dorsal columns. Histoautoradiography and immunohistochemistry have clarified the developmental steps in the complex specification of DRG neurons (10–13), but it is unknown how frataxin deficiency impacts this elaborately tooled process. Total absence of frataxin in genetically engineered mice is lethal to the embryo (14), but it must still be determined whether spinal cord and DRG development proceeds normally when frataxin is only low rather than totally absent. It is impossible to establish from autopsy tissues of patients who succumbed to FA after years of illness how neural crest abnormalities cause hypoplasia of spinal cord and DRG. The current study used quantitative measurements to establish a possible correlation of the cross-sectional area of the thoracic spinal cord and neuronal sizes in lumbosacral DRG. The results strongly support the conclusion that these neural tissues are small because of hypoplasia rather than atrophy and that their developmental delay occurs independently. MATERIALS AND METHODS The table summarizes clinical and genetic data of 24 FA patients. The cases were selected from among 47 FA autopsies because the harvested tissues included thoracic spinal cord and lumbosacral DRG. One patient (FA4) with early onset and short survival was a compound heterozygote with a guanine-adenine-adenine (GAA) trinucleotide expansion of 730 on 1 allele and a deletion (c.11_12delTC) on the other allele. We considered inclusion of this patient justified because clinical features were similar to those of the other FA patients with early onset and short duration. The neuropathology was identical to a patient with long biallelic GAA expansions (FA1; GAA1, 1016; GAA2, 1016). Disease onset in 2 patients was delayed to 32 and 34 years (FA23 and FA24, Table), and their deaths occurred at 87 and 77 years, respectively. In FA23, FA was not diagnosed during life although the patient was ataxic. The diagnosis was assured, however, by characteristic neuropathology and genetic testing of DNA extracted from frozen cerebellar cortex. In homozygous FA patients, ages of onset and disease duration correlated inversely with the shorter GAA trinucleotide repeat expansion (Table). The cause of death in 19 patients was cardiomyopathy (79%). Other causes of death were cachexia (FA3), myoglobinuria and renal failure (FA16), myocardial infarction (FA17), brain hemorrhage (FA20), and coronary atherosclerosis (FA23). The hearts of all patients were available for detailed histological study, and iron-reactive inclusion in cardiomyocytes were detectable in 23 of 24 cases (96%). Although the spinal cord in FA is thin at all levels, we elected to measure the cross-sectional areas at the midthoracic level because they change little between the second and twelfth segments (15), and long-tract disease in FA is most distinct (Fig. 1). Measurements of neuronal areas in lumbosacral rather than thoracic DRG were preferred because of the vast sensory input from legs and perineum, and the known correlation of DRG neuronal size and the extent of the innervated region (16). Transverse sections of thoracic spinal cord were embedded in paraffin, sectioned at 6-µm thickness, and stained by immunohistochemistry with an antibody to myelin basic protein (MBP). Sections of DRG were stained with hematoxylin and eosin. Twelve adult control cases with ages of death from 41 to 90 years (mean ± SD: 65 ± 15 years) were made available by the autopsy services of Veterans Affairs Medical Center in Albany, NY, and Albany Medical College. Inclusion criteria were absence of neurological disease of any kind and availability of the midthoracic spinal cord and lumbar DRG. In some older archival control cases, paraffin blocks were no longer available, and measurements of the spinal cord were made on sections stained with hematoxylin and eosin or the Luxol fast blue-periodic acid-Schiff method for myelin. In selected cases of FA and normal controls, 1.5-cm-long segments of upper lumbar spinal cord were trimmed longitudinally to represent DR entry zones. Paraffin sections were immunostained for phosphorylated neurofilament protein, peripheral myelin protein 22 (PMP22), and proteolipid protein (PLP). FIGURE 1 Open in new tabDownload slide Area measurements of the thoracic spinal cord and DRG neurons. (a–c) FA (FA18, Table); (d–f) normal control (man, age of death, 62 years). (a, d) Immunostain of MBP; (b, c, e, f) hematoxylin and eosin. The transverse section of the thoracic spinal cord in FA (a) shows the overall reduction in size and the characteristic symmetrical loss of myelinated fibers in dorsal columns, dorsal spinocerebellar, and lateral corticospinal tracts. After outlining the limits of the spinal cord in FA (a) and the control (d) with the spline feature of the program, cross-sectional areas were measured as 19.11 and 36.53 mm2, respectively. DRG neurons with distinctly visible nucleoli were similarly outlined and measured (b, c, e, f). When the satellite layer and neuronal plasma membrane were separated due to postmortem artifact, the outline was placed midway between satellite cells and the neuron (18). In FA, neurons were generally smaller (b, c). The section shown in (c) reveals residual nodules (arrows). In (b), a neuron is undergoing active invasion by satellite cells or monocytes (arrow). Scale bars: a, d, 1 mm; b, c, e, f, 20 µm. FIGURE 1 Open in new tabDownload slide Area measurements of the thoracic spinal cord and DRG neurons. (a–c) FA (FA18, Table); (d–f) normal control (man, age of death, 62 years). (a, d) Immunostain of MBP; (b, c, e, f) hematoxylin and eosin. The transverse section of the thoracic spinal cord in FA (a) shows the overall reduction in size and the characteristic symmetrical loss of myelinated fibers in dorsal columns, dorsal spinocerebellar, and lateral corticospinal tracts. After outlining the limits of the spinal cord in FA (a) and the control (d) with the spline feature of the program, cross-sectional areas were measured as 19.11 and 36.53 mm2, respectively. DRG neurons with distinctly visible nucleoli were similarly outlined and measured (b, c, e, f). When the satellite layer and neuronal plasma membrane were separated due to postmortem artifact, the outline was placed midway between satellite cells and the neuron (18). In FA, neurons were generally smaller (b, c). The section shown in (c) reveals residual nodules (arrows). In (b), a neuron is undergoing active invasion by satellite cells or monocytes (arrow). Scale bars: a, d, 1 mm; b, c, e, f, 20 µm. Host, clonality, supplier, catalogue number, antigen retrieval method, and final protein concentration of antibodies were as follows: MBP, mouse monoclonal (IgG), Covance, Emeryville, CA; SMI-94R, 80% ethanol overnight at 4 °C, ascites fluid (protein not assayed); PMP22, rabbit polyclonal (IgG), Abcam, Cambridge, MA, ab15506, 80% ethanol overnight at 4 °C, 0.4 μg/ml; PLP, rabbit polyclonal (IgG), Abcam, ab105784, 80% ethanol overnight at 4 °C, followed by 0.01M citric acid-sodium citrate buffer (pH 6) at 95 °C for 20 minutes, 2 μg/ml; phosphorylated neurofilament protein, mouse monoclonal (IgG), Sternberger Monoclonals, Inc., Lutherville, MD (now Covance), SMI-31, 0.01 M citric acid-sodium citrate buffer (pH 6) at 95 °C for 20 min, ascites fluid (protein not assayed). A summary of the immunohistochemical procedure is available (7). Image capture and measurements of the cross-sectional areas of the spinal cord and perikaryal area of DRG neurons were made with AxioVision software (version 4.8) (Carl Zeiss, Oberkochen, Germany). Thoracic spinal cord areas were expressed as mm2 and neuronal areas as µm2. Selection of DRG neurons for measurement was based on the presence of distinct nuclei and nucleoli (17). DRG exhibit a common autolytic artifact, namely, separation of the neuronal plasma membrane from the cytoplasmic borders of surrounding satellite cells. In an effort to compensate for this artifact, the measuring spline was drawn midway between satellite cells and the neuronal surface, as practiced by Ohta et al (18). To prevent duplicate measurement of DRG neurons, the specimen stage of the microscope was moved in register across the section. Statistical Analysis Comparison of FA patients and normal controls was by Student t-test with values expressed as mean ± SD. Correlation between thoracic spinal cord area and DRG neuronal size and GAA repeat expansion were assessed by regression. To model the thoracic spinal cord area as a function of age, we used non-linear regression (OriginPro 2016, Origin Lab Corporation, Northampton, MA) under the assumption that the area is 11.56 mm2 at birth (19) and subsequent growth is exponential, reaching a constant fully mature size. The equation was therefore S(t) = 11.56 + (Smature) (1−e−kt) where S is the thoracic spinal cord area as a function of time (t) in years; Smature the increase in the area of the spinal cord from birth to maturity; and k the growth constant. Thoracic spinal cord areas at younger ages were obtained from the literature (15, 19) and the growth rate constant was assumed to be equal in both groups. RESULTS Figure 1 shows composite photomicrographs of thoracic spinal cord and DRG of an FA patient (FA18, Table) (Fig. 1a–c) and a normal control (man, age of death: 62 years) (Fig. 1d–f). The stain for MBP shows the small size of the cord and the severe symmetrical depletion of myelinated fibers in dorsal columns, dorsal spinocerebellar, and lateral corticospinal tracts (Fig. 1a). The section of the FA case also shows anteroposterior flattening of the thoracic spinal cord when compared with the normal control (Fig. 1d). Neurons in the DRG of FA (Fig. 1b, c) are generally smaller in comparison with normal DRG nerve cells (Fig. 1e, f). In FA, DRG are hypercellular, residual nodules are frequent (Fig. 1c, arrows); and a cluster of small cells suggests active invasion of a neuron by mononuclear cells, satellite cells, or both (Fig. 1b, arrow). Figure 2 shows a composite of a DR entry zone in FA (FA22, Table) and a control case after immunohistochemical staining for phosphorylated neurofilament protein (Fig. 2a, b), PMP22 (Fig. 2c, d), and PLP (Fig. 2e, f). Reaction product of phosphorylated neurofilament protein shows transition of thin axons from the DR into the central nervous system (CNS) of the spinal cord in FA (Fig. 2a, arrow), but the sharp demarcation of relatively large axons in normal DR is absent (Fig. 2b, arrow). Similarly, the DR entry zone in FA lacks the abrupt ending of PMP22 (Fig. 2c, arrow), which characterizes the normal DR (Fig. 2d, arrow). Reaction product of the CNS-specific protein, PLP, mirrors the disorganized transition of PMP22-positive DR fibers in FA (Fig. 2e, arrow), and stands out against the sharply demarcated CNS myelin in the normal DR (Fig. 2f, arrow). FIGURE 2 Open in new tabDownload slide DR entry zone in FA and a normal control. (a, c, e) FA (FA22, Table); (b, d, f) normal control (woman, age of death, 42 years). Adjacent sections were immunostained with antibodies to phosphorylated neurofilament protein (a, b); PMP22 (c, d) or PLP (e, f). The cranial direction of the spinal cord is to the right. In FA, the DR contains an abundance of thin axons (a), but the transition from relatively thick axons of the peripheral nervous system (PNS) to thin CNS axons is indistinct (a, arrow). In the normal control, the transition from PNS to CNS axons occurs over a very narrow segment of the DR (b, arrow), and larger axons are restricted to the PNS portion of the root. In normal DR, PMP22 immunoreactivity ends abruptly (d, arrow) whereas in FA, some PMP22-reactive fibers can be traced across the entire DR entry zone (c, arrow). In the normal control, PLP-reactive myelin sheaths begin at a considerable distance (∼200 µm) before entering the spinal cord proper, and the transition to CNS myelin is sharp (f, arrow). In FA, PLP immunoreactivity in the CNS portion of the DR is sparse, and the transition from PNS to CNS territory is irregular (e, arrow). Bars: 100 µm. FIGURE 2 Open in new tabDownload slide DR entry zone in FA and a normal control. (a, c, e) FA (FA22, Table); (b, d, f) normal control (woman, age of death, 42 years). Adjacent sections were immunostained with antibodies to phosphorylated neurofilament protein (a, b); PMP22 (c, d) or PLP (e, f). The cranial direction of the spinal cord is to the right. In FA, the DR contains an abundance of thin axons (a), but the transition from relatively thick axons of the peripheral nervous system (PNS) to thin CNS axons is indistinct (a, arrow). In the normal control, the transition from PNS to CNS axons occurs over a very narrow segment of the DR (b, arrow), and larger axons are restricted to the PNS portion of the root. In normal DR, PMP22 immunoreactivity ends abruptly (d, arrow) whereas in FA, some PMP22-reactive fibers can be traced across the entire DR entry zone (c, arrow). In the normal control, PLP-reactive myelin sheaths begin at a considerable distance (∼200 µm) before entering the spinal cord proper, and the transition to CNS myelin is sharp (f, arrow). In FA, PLP immunoreactivity in the CNS portion of the DR is sparse, and the transition from PNS to CNS territory is irregular (e, arrow). Bars: 100 µm. The adult thoracic spinal cord area in FA (death at >24 years, Table) is 23.83 ± 3.68 mm2 (mean ± SD), which is significantly lower than the mean in normal adults (42.19 ± 10.35, including data from the literature) (p < 0.001) (Table). Figure 3 shows a regression analysis of thoracic spinal cord areas vs age of death (Table). The curve for FA was constructed under the assumptions that newborns destined to develop FA have the same thoracic spinal cord areas as normal newborns (11.56 mm2 [19]) and that the area increases exponentially to a final size. The regression curve for FA shows that the spinal cord fails to achieve normal adult size and remains stable over the entire disease duration or at least does not significantly decrease in those who die at more advanced age. In normal controls, the spinal cord grows rapidly and reaches its final size at the approximate age of 25 years. The growth phase of these curves should be interpreted with caution as we have little information aside from the literature on the rate of growth of the spinal cord in normal humans, and no information at all on FA. FIGURE 3 Open in new tabDownload slide Comparison of growth curves of the thoracic spinal cord areas in FA and normal controls. Red triangles and red line, FA; blue squares and blue line, normal controls. Values taken from the literature are labeled as “Lit” (normal controls only). Curves were fitted by a non-linear regression program (OriginPro 2016) under the assumption that the thoracic spinal cord area is the same for FA and normal controls at birth, and subsequent growth is exponential until reaching a fully mature size. Thereafter, size is assumed to be constant without increasing or decreasing. The underlying equation is S(t) = 11.56 + (Smature)(1−e−kt), where S is the thoracic spinal cord area as a function of time t (in years); 11.56 is the cord size in mm2 at t = 0; Smature is the increase in size from birth to maturity; and k is the rate constant of growth. k has units of inverse time. 1/k will be the time constant for growth and 0.693/k is the half time of growth. Both FA and normal controls show rapid growth of the spinal cord over the first 10 years of life, but FA fails to achieve normal adult dimensions. The spinal cord in FA remains hypoplastic throughout life and does not undergo progressive thinning as a function of age. FIGURE 3 Open in new tabDownload slide Comparison of growth curves of the thoracic spinal cord areas in FA and normal controls. Red triangles and red line, FA; blue squares and blue line, normal controls. Values taken from the literature are labeled as “Lit” (normal controls only). Curves were fitted by a non-linear regression program (OriginPro 2016) under the assumption that the thoracic spinal cord area is the same for FA and normal controls at birth, and subsequent growth is exponential until reaching a fully mature size. Thereafter, size is assumed to be constant without increasing or decreasing. The underlying equation is S(t) = 11.56 + (Smature)(1−e−kt), where S is the thoracic spinal cord area as a function of time t (in years); 11.56 is the cord size in mm2 at t = 0; Smature is the increase in size from birth to maturity; and k is the rate constant of growth. k has units of inverse time. 1/k will be the time constant for growth and 0.693/k is the half time of growth. Both FA and normal controls show rapid growth of the spinal cord over the first 10 years of life, but FA fails to achieve normal adult dimensions. The spinal cord in FA remains hypoplastic throughout life and does not undergo progressive thinning as a function of age. The nerve cell area in DRG of FA (1362 ± 442 µm2 [mean ± SD]) is significantly lower than normal (2004 ± 233 µm2) (p < 0.001). Figure 4 shows a display of the mean area of neurons in DRG of FA and normal controls plotted against the area of the thoracic spinal cord in matching specimens (FA cases from the Table). There is no correlation in FA or normal controls, implying that delayed growth of DRG neurons and spinal cord in FA proceeds independently over time. Spinal cord areas and neuronal sizes in DRG of FA patients did not correlate significantly with the GAA trinucleotide repeats on either allele (Table). FIGURE 4 Open in new tabDownload slide Correlation of DRG neuronal sizes and thoracic spinal cord area in FA and normal controls. Red triangles and red line, FA (labeled according to the Table); blue squares and blue line, normal controls. Spinal cord areas and mean nerve cell sizes in DRG are lower in FA than in normal controls but linear regression shows no significant correlation between DRG and spinal cord measurements. Neither slope is significantly different from zero (p = 0.29; R2 = 0.11 for normal and p = 0.47; R2 = 0.02 for FA). FIGURE 4 Open in new tabDownload slide Correlation of DRG neuronal sizes and thoracic spinal cord area in FA and normal controls. Red triangles and red line, FA (labeled according to the Table); blue squares and blue line, normal controls. Spinal cord areas and mean nerve cell sizes in DRG are lower in FA than in normal controls but linear regression shows no significant correlation between DRG and spinal cord measurements. Neither slope is significantly different from zero (p = 0.29; R2 = 0.11 for normal and p = 0.47; R2 = 0.02 for FA). DISCUSSION Hypoplasia or Atrophy of Spinal Cord and DRG Neurons in FA FA is commonly called a “neurodegenerative” disease, implying that the nervous system at one time during the patient’s life is normal and then undergoes “atrophy”. In the absence of information about the sizes of spinal cord and DRG neurons before disease onset, these terms may no longer be justified. Length and volume of the human spinal cord increase massively between birth and maturity, and most of the accrual is due to white matter (19). Pooling the meticulous planimetric measurements of the cross-sectional area of the thoracic spinal cord (segment VI) by Stilling at the ages of 1, 2, 25, 35, and 45 years (15), the combined measurements of thoracic spinal gray and white matter in newborns (19), and measurements in this laboratory allowed the construction of a growth curve of the normal spinal cord that can be compared with FA (Fig. 3). Based on the equation shown in the legend to Figure 3 for normal spinal cords, the expected thoracic cross-sectional areas in our FA patients succumbing to early deaths (FA1 and FA4 in the Table) should be 40.5 mm2 (death at 10 years) and 40.9 mm2 (death at 11 years), respectively. The measurements, however, were 23.7 and 32.0 mm2. It is likely that planimetry of camera clara or lucida drawings practiced by the earlier authors (15, 19) was less accurate than the quantitative method described here. The areas determined in the 2 rapidly fatal cases of FA by our method did not differ significantly from those in adult FA patients (Table) who died between the ages of 23 and 87. These observations support the interpretation that spinal cord enlargement in FA stops prematurely after growth at either a normal or a reduced rate, representing of hypoplasia. The exacting measurements of neuronal diameters in the first sacral DRG of humans by Ohta et al (18) revealed rapid growth of nerve cell sizes between the ages of 0.5 and 12 years with little change thereafter into advanced age. Our measurements of neuronal areas in normal controls are not directly comparable to data published by Ohta et al due to technical differences and a less selective harvesting of lumbosacral DRG. Nevertheless, the significantly lower neuronal sizes in DRG of FA patients undergo little change between the ages of 10 and 87 years (Table). This result implies that hypoplasia also affects DRG neurons in FA. In this context, inflammatory destruction of DRG neurons (7) may be a superimposed process and ultimately less important than dysfunction due to life-long hypoplasia. Previous systematic measurements of DR nerve fibers in FA and a comparison with ventral roots also militated against atrophy (20). Histograms published by Koeppen et al confirmed Friedreich’s original report (1) that DR lacked large myelinated fibers (20). Surprisingly, the overall degree of myelination of thin DR axons was significantly higher (20). Only 2 groups of authors have examined the transition of DR fibers from the PNS to the CNS in FA (21, 22). Hughes et al (22) traced DR fibers into the spinal cord and concluded that they reached the gray matter of the dorsal horns while few entered the dorsal columns. In contrast, Jitpimolmard et al favored regeneration of thin fibers to account for their abnormal abundance (21). Our data (Fig. 2a) support the former conclusion, i.e. that DR axons traverse the entry zone though the abrupt transition from thicker peripheral axons to thin centrally projecting fibers is absent. Several other observations suggest that a developmental abnormality of DR accounts for the fiber depletion in the spinal cord. It is likely that the disorganization of the DR entry zone reflects incorrect signaling to growing axons during development. An incidental observation in the 2 very young patients (FA1 and FA4, Table) was the presence in their DR of bizarre expansions that reacted with antibodies to S100 and glial fibrillary acidic protein (not illustrated). This observation may indicate an abnormal Schwann cell response in FA that is no longer detectable in adult cases of FA. The spinal cord areas normally occupied by the dorsal columns in FA are clearly recognizable although there is subtotal loss of myelinated fibers. Gliosis characteristically occurs in sheaf-like bundles, suggesting that DRG at one time successfully projected axons through DR and dorsal columns to reach the gracile and cuneate nuclei. The presence of gliosis, however, is consistent with superimposed degeneration before the spinal cord reaches its final transverse size (Fig. 3). One of the most dramatic lesions in the spinal cord of FA is the total or subtotal lack of large nerve cells in the dorsal nuclei of Clarke columns at thoracic and upper lumbar spinal levels. While interpretable as transsynaptic degeneration, the timing of this deafferentation in FA is unknown. Even in the 2 very young patients with FA, the neurons in the dorsal nuclei were greatly reduced in number (FA1, Table) or entirely absent (FA4, Table). It is likely that these normally rather large nerve cells failed to survive due to lack of innervation from DR collaterals that normally occurs between 14 and 17.5 weeks of gestation (23). Muscle spindle afferents make monosynaptic contacts with anterior horn cells, and areflexia is a common observation in FA. Motor neurons of the spinal cord, however, are not seriously affected by FA though ventral roots may show more numerous clusters of small fibers compared with normal ventral roots (20). There is also significant electrodiagnostic evidence that muscles undergo denervation in FA (24). In situ hybridization of mouse embryos with digoxigenin-labeled antisense riboprobes first detects frataxin transcripts in DRG at the gestational ages of 14.5–16.5 days (25), at which time the spinal cord also shows frataxin expression. Assuming that DRG and spinal cord are similar in the human embryo, high prenatal frataxin expression may make these tissues vulnerable later in life, and the onset of developmental delay may have to be placed into prenatal life. The Functional Consequences of Hypoplastic DRG and Spinal Cord in FA, and Therapeutic Implications The described observations suggest that the term “onset” of FA may have to be redefined. Some parents report that their children who are ultimately affected by the disease did not develop motor skills at the same pace as their unaffected siblings. It remains to be established how clinical onset occurs on a background of neural hypoplasia. Many clinicians charged with the care of FA patients rely on the length of GAA trinucleotide repeats, and notably the shorter GAA expansion, to set a prognosis. Statistical analysis, however, reveals no correlation between the lengths of GAA trinucleotide repeats and spinal cord area or DRG neuronal size (R2 < 0.1). This observation may be surprising in light of the inverse correlation between the shorter GAA expansion and clinical onset of FA (Table). Very short GAA trinucleotide repeat expansions convey a benign course and long survival, such as in FA23 (Table), and cardiomyopathy in such cases may be entirely absent. Hypoplasia of spinal cord and DRG do not appear to conform to a gradient of GAA trinucleotide expansion. It is likely that even modest reductions of frataxin are detrimental to the development of spinal cord and DRG. In contrast to DRG and spinal cord, thinning of the DN in FA represents true atrophy (26). The DN may remain intact in cases of rapidly fatal FA cardiomyopathy (27); and it is reasonable to correlate the onset of the cerebellar manifestations of FA with a threshold level of neuronal loss. Onset of proprioceptive and superficial sensory loss may coincide with beginning inflammatory infiltration of DRG (7). It is uncertain how upper motor neuron weakness correlates with progressive atrophy of the corticospinal tracts. Loss of Betz cells is common in FA, but these very large neurons of the frontal cortex account for only a small minority of fibers in the corticospinal tracts. The recognition of hypoplasia in the pathogenesis of FA may redirect therapy. At this time, onset is defined by clinical criteria, and disability is quantified by one or more widely used ataxia rating scales. Clinical trials with drugs generally involve symptomatic FA patients in a specific age group. FA is an uncommon disease (28), but newborn screening for the FA mutation may be justified as the downstream effects of frataxin deficiency are better defined. The task of preventing growth arrest of spinal cord and DRG evolving between birth and maturity will differ from therapy in established FA patients. It seems reasonable to propose that drugs raising systemic frataxin levels should be used mainly for presymptomatic FA patients. TABLE Patient Data No. . Sex . Age of onset (years) . Age at death (years) . Disease duration (years) . GAA1 . GAA2 . Spinal cord area (mm2) . Mean neuronal area (µm2) . FA1 M 2 10 8 1016 1016 23.66 1452 FA2 F 2 33 31 455 604 29.06 1296 FA3 F 5 25 20 800 1100 25.98 1189 FA4 M 5 11 6 730 deletion 31.96 1402 FA5 M 7 34 27 1114 1114 18.94 1488 FA6 M 7 35 28 750 1000 24.54 1058 FA7 F 8 28 20 559 709 23.56 1616 FA8 F 8 23 15 528 777 21.75 1530 FA9 F 9 26 17 690 850 24.44 1571 FA10 M 9 33 24 925 925 29.16 1453 FA11 F 10 25 15 750 850 24.05 1183 FA12 F 10 24 14 700 950 27.32 1198 FA13 M 10 38 28 249 934 19.23 1725 FA14 F 10 47 37 613 613 19.36 1029 FA15 F 11 42 31 761 990 17.15 1238 FA16 F 12 24 12 740 910 26.42 1155 FA17 M 14 50 36 400 750 27.37 1471 FA18 F 15 69 54 568 568 19.11 1141 FA19 M 16 46 30 599 599 23.76 1418 FA20 F 17 58 41 566 566 24.33 1252 FA21 F 17 50 33 512 1122 19.16 1241 FA22 F 18 67 49 621 766 28.99 1715 FA23 M 32 87 55 120 170 26.34 1917 FA24 F 34 77 43 466 841 24.36 950 Mean ± SD 12 ± 7.8 40.1 ± 20.1 28.1 ± 13.8 635 ± 222 814 ± 225 24.17 ± 3.89 1362 ± 242 No. . Sex . Age of onset (years) . Age at death (years) . Disease duration (years) . GAA1 . GAA2 . Spinal cord area (mm2) . Mean neuronal area (µm2) . FA1 M 2 10 8 1016 1016 23.66 1452 FA2 F 2 33 31 455 604 29.06 1296 FA3 F 5 25 20 800 1100 25.98 1189 FA4 M 5 11 6 730 deletion 31.96 1402 FA5 M 7 34 27 1114 1114 18.94 1488 FA6 M 7 35 28 750 1000 24.54 1058 FA7 F 8 28 20 559 709 23.56 1616 FA8 F 8 23 15 528 777 21.75 1530 FA9 F 9 26 17 690 850 24.44 1571 FA10 M 9 33 24 925 925 29.16 1453 FA11 F 10 25 15 750 850 24.05 1183 FA12 F 10 24 14 700 950 27.32 1198 FA13 M 10 38 28 249 934 19.23 1725 FA14 F 10 47 37 613 613 19.36 1029 FA15 F 11 42 31 761 990 17.15 1238 FA16 F 12 24 12 740 910 26.42 1155 FA17 M 14 50 36 400 750 27.37 1471 FA18 F 15 69 54 568 568 19.11 1141 FA19 M 16 46 30 599 599 23.76 1418 FA20 F 17 58 41 566 566 24.33 1252 FA21 F 17 50 33 512 1122 19.16 1241 FA22 F 18 67 49 621 766 28.99 1715 FA23 M 32 87 55 120 170 26.34 1917 FA24 F 34 77 43 466 841 24.36 950 Mean ± SD 12 ± 7.8 40.1 ± 20.1 28.1 ± 13.8 635 ± 222 814 ± 225 24.17 ± 3.89 1362 ± 242 Patients are listed in order of age of onset, also showing sex, age of death, disease duration, GAA1 and GAA2, thoracic spinal cord area in mm2, and mean area of DRG neurons in µm2. In 12 adult locally procured controls (F, 2; M, 10), the age of death was 65 ± 15 years and ranged from 41 to 90 years. The thoracic spinal cord area in these cases is 37.3 ± 3.6 mm2. In matching DRG, the neuronal area is 2004 ± 233 µm2 (mean ± SD). The differences between FA and controls are significant at p < 0.001 (spinal cord and DRG neuronal areas). Data are expressed as mean ± SD. F, female; FA, Friedreich ataxia; DRG, dorsal root ganglia; GAA, guanine-adenine-adenine trinucleotide expansion; M, male; SD, standard deviation. Open in new tab TABLE Patient Data No. . Sex . Age of onset (years) . Age at death (years) . Disease duration (years) . GAA1 . GAA2 . Spinal cord area (mm2) . Mean neuronal area (µm2) . FA1 M 2 10 8 1016 1016 23.66 1452 FA2 F 2 33 31 455 604 29.06 1296 FA3 F 5 25 20 800 1100 25.98 1189 FA4 M 5 11 6 730 deletion 31.96 1402 FA5 M 7 34 27 1114 1114 18.94 1488 FA6 M 7 35 28 750 1000 24.54 1058 FA7 F 8 28 20 559 709 23.56 1616 FA8 F 8 23 15 528 777 21.75 1530 FA9 F 9 26 17 690 850 24.44 1571 FA10 M 9 33 24 925 925 29.16 1453 FA11 F 10 25 15 750 850 24.05 1183 FA12 F 10 24 14 700 950 27.32 1198 FA13 M 10 38 28 249 934 19.23 1725 FA14 F 10 47 37 613 613 19.36 1029 FA15 F 11 42 31 761 990 17.15 1238 FA16 F 12 24 12 740 910 26.42 1155 FA17 M 14 50 36 400 750 27.37 1471 FA18 F 15 69 54 568 568 19.11 1141 FA19 M 16 46 30 599 599 23.76 1418 FA20 F 17 58 41 566 566 24.33 1252 FA21 F 17 50 33 512 1122 19.16 1241 FA22 F 18 67 49 621 766 28.99 1715 FA23 M 32 87 55 120 170 26.34 1917 FA24 F 34 77 43 466 841 24.36 950 Mean ± SD 12 ± 7.8 40.1 ± 20.1 28.1 ± 13.8 635 ± 222 814 ± 225 24.17 ± 3.89 1362 ± 242 No. . Sex . Age of onset (years) . Age at death (years) . Disease duration (years) . GAA1 . GAA2 . Spinal cord area (mm2) . Mean neuronal area (µm2) . FA1 M 2 10 8 1016 1016 23.66 1452 FA2 F 2 33 31 455 604 29.06 1296 FA3 F 5 25 20 800 1100 25.98 1189 FA4 M 5 11 6 730 deletion 31.96 1402 FA5 M 7 34 27 1114 1114 18.94 1488 FA6 M 7 35 28 750 1000 24.54 1058 FA7 F 8 28 20 559 709 23.56 1616 FA8 F 8 23 15 528 777 21.75 1530 FA9 F 9 26 17 690 850 24.44 1571 FA10 M 9 33 24 925 925 29.16 1453 FA11 F 10 25 15 750 850 24.05 1183 FA12 F 10 24 14 700 950 27.32 1198 FA13 M 10 38 28 249 934 19.23 1725 FA14 F 10 47 37 613 613 19.36 1029 FA15 F 11 42 31 761 990 17.15 1238 FA16 F 12 24 12 740 910 26.42 1155 FA17 M 14 50 36 400 750 27.37 1471 FA18 F 15 69 54 568 568 19.11 1141 FA19 M 16 46 30 599 599 23.76 1418 FA20 F 17 58 41 566 566 24.33 1252 FA21 F 17 50 33 512 1122 19.16 1241 FA22 F 18 67 49 621 766 28.99 1715 FA23 M 32 87 55 120 170 26.34 1917 FA24 F 34 77 43 466 841 24.36 950 Mean ± SD 12 ± 7.8 40.1 ± 20.1 28.1 ± 13.8 635 ± 222 814 ± 225 24.17 ± 3.89 1362 ± 242 Patients are listed in order of age of onset, also showing sex, age of death, disease duration, GAA1 and GAA2, thoracic spinal cord area in mm2, and mean area of DRG neurons in µm2. In 12 adult locally procured controls (F, 2; M, 10), the age of death was 65 ± 15 years and ranged from 41 to 90 years. The thoracic spinal cord area in these cases is 37.3 ± 3.6 mm2. In matching DRG, the neuronal area is 2004 ± 233 µm2 (mean ± SD). The differences between FA and controls are significant at p < 0.001 (spinal cord and DRG neuronal areas). Data are expressed as mean ± SD. F, female; FA, Friedreich ataxia; DRG, dorsal root ganglia; GAA, guanine-adenine-adenine trinucleotide expansion; M, male; SD, standard deviation. Open in new tab Funding: New York State Department of Health, Albany, NY, USA; Friedreich’s Ataxia Research Alliance, Downingtown, PA, and Neurochemical Research, Inc., Glenmont, NY. Disclosure/conflict of interest: The authors have no duality or conflicts of interest to declare. ACKNOWLEDGMENTS We thank the families who agreed to donate autopsy tissue for FA research. Many pathologists agreed to complete detailed autopsies on FA patients though death in remote areas of the United States often made this task very difficult. The work described in this article was completed in the research laboratories of the Veterans Affairs Medical Center in Albany, NY, USA. Michigan State University, East Lansing, MI, USA made available a copy of the classic 1859 treatise on the structure of the spinal cord by Stilling (15). REFERENCES 1 Friedreich N. Ueber Ataxie mit besonderer Berücksichtigung der hereditären Formen . Nachtrag. Virchows Arch Pathol Anat Physiol Klin Med 1877 ; 70 : 140 – 52 Google Scholar Crossref Search ADS WorldCat 2 Koeppen AH , Neuropathology of the inherited ataxias. In: Manto M-U Pandolfo M, eds., The Cerebellum and its Disorders , Cambridge University Press , Cambridge 2002 : 387 – 405 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 3 Vassilopoulos D Spengos M Scarpalezos S. Étude radiologique de la colonne vertébrale cervicale dans certaines maladies dégéneratives neurologiques . J Radiol Electrol 1977 ; 58 : 183 – 6 Google Scholar PubMed OpenURL Placeholder Text WorldCat 4 Wessel K Schroth G Diener HC, et al. . Significance of MRI-confirmed atrophy of the cranial spinal cord in Friedreich’s ataxia . Eur Arch Psychiatry Neurol Sci 1989 ; 238 : 225 – 30 Google Scholar Crossref Search ADS PubMed WorldCat 5 Mascalchi M Salvi F Piacentini S, et al. . Friedreich’s ataxia: MR findings involving the cervical portion of the spinal cord . AJR Am J Roentgenol 1994 ; 163 : 187 – 91 Google Scholar Crossref Search ADS PubMed WorldCat 6 Chevis CF da Silva CB D’Abreu A, et al. . Spinal cord atrophy correlates with disability in Friedreich’s ataxia . Cerebellum 2013 ; 12 : 43 – 7 Google Scholar Crossref Search ADS PubMed WorldCat 7 Koeppen AH Ramirez RL Becker AB, et al. . Dorsal root ganglia in Friedreich ataxia: satellite cell proliferation and inflammation . Acta Neuropathol Commun 2016 ; 4 : 46 Google Scholar Crossref Search ADS PubMed WorldCat 8 Pannese E. The histogenesis of the spinal ganglia . Adv Anat Embryol Cell Biol 1974 ; 47 : 7 – 97 Google Scholar PubMed OpenURL Placeholder Text WorldCat 9 Maro GS Vermeren M Voiculescu O, et al. . Neural crest boundary cap cells constitute a source of neuronal and glial cells of the PNS . Nat Neurosci 2004 ; 7 : 930 – 8 Google Scholar Crossref Search ADS PubMed WorldCat 10 Lawson SN Caddy KWT Biscoe TJ. Development of rat dorsal root ganglion neurons . Cell Tissue Res 1974 ; 153 : 399 – 413 Google Scholar Crossref Search ADS PubMed WorldCat 11 Lawson SN Biscoe TJ. Development of mouse dorsal root ganglia: an autoradiographic and quantitative study . J Neurocytol 1979 ; 8 : 265 – 74 Google Scholar Crossref Search ADS PubMed WorldCat 12 Kitao Y Robertson B Kudo M, et al. . Neurogenesis of subpopulations of rat lumbar dorsal root ganglion neurons including neurons projecting to the dorsal column nuclei . J Comp Neurol 1996 ; 371 : 249 – 57 Google Scholar Crossref Search ADS PubMed WorldCat 13 Raible DW Ungos JM. Specification of sensory neuron cell fate from the neural crest . Adv Exp Med Biol 2006 ; 589 : 170 – 80 Google Scholar PubMed OpenURL Placeholder Text WorldCat 14 Cossée M Puccio H Gansmuller A, et al. . Inactivation of the Friedreich ataxia mouse gene leads to early embryonic lethality without iron accumulation . Hum Mol Genet 2000 ; 9 : 1219 – 26 Google Scholar Crossref Search ADS PubMed WorldCat 15 Stilling B. Neue Untersuchungen ueber den Bau des Rückenmarks . Heinrich Hotop . Cassel 1859 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 16 Lieberman AR , Sensory ganglia. In: Landon DN, editor. The peripheral nerve . Chapman and Hall , London 1976 : 188 - 278 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 17 Aherne W. Some morphometric methods for the central nervous system . J Neurol Sci 1975 ; 24 : 221 – 41 Google Scholar Crossref Search ADS PubMed WorldCat 18 Ohta M Offord K Dyck PJ. Morphometric evaluation of first sacral ganglia of man . J Neurol Sci 1974 ; 22 : 73 – 82 Google Scholar Crossref Search ADS PubMed WorldCat 19 Lassek AM Rasmussen GL. A quantitative study of the newborn and adult spinal cords of man . J Comp Neurol 1938 ; 69 : 371 – 9 Google Scholar Crossref Search ADS WorldCat 20 Koeppen AH Morral JA Davis AN, et al. . The dorsal root ganglion in Friedreich’s ataxia . Acta Neuropathol 2009 ; 118 : 763 – 76 Google Scholar Crossref Search ADS PubMed WorldCat 21 Jitpimolmard S Small J King RHM, et al. . The sensory neuropathy of Friedreich’s ataxia: an autopsy study of a case with prolonged survival . Acta Neuropathol 1993 ; 86 : 29 – 35 Google Scholar Crossref Search ADS PubMed WorldCat 22 Hughes JT Brownell B Hewer RL. The peripheral sensory pathway in Friedreich’s ataxia. An examination by light and electron microscopy of the posterior nerve roots, posterior root ganglia, and peripheral sensory nerves in cases of Friedreich’s ataxia . Brain 1968 ; 91 : 803 – 18 Google Scholar Crossref Search ADS PubMed WorldCat 23 Altman J Bayer SA, Development of the Human Spinal Cord. An Interpretation Based on Experimental Studies in Animals . Oxford : Oxford University Press 2001 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 24 Caruso G Santoro L Perretti A, et al. . Friedreich’s ataxia: electrophysiological and histological findings . Acta Neurol Scand 1983 ; 67 : 26 – 40 Google Scholar Crossref Search ADS PubMed WorldCat 25 Jiralerspong S Liu Y Montermini L, et al. . Frataxin shows developmentally regulated tissue-specific expression in the mouse embryo . Neurobiol Dis 1997 ; 4 : 103 – 13 Google Scholar Crossref Search ADS PubMed WorldCat 26 Koeppen AH Ramirez RL Becker AB, et al. . Friedreich ataxia: failure of GABA-ergic and glycinergic synaptic transmission in the dentate nucleus . J Neuropathol Exp Neurol 2015 ; 74 : 166 – 76 Google Scholar Crossref Search ADS PubMed WorldCat 27 Koeppen AH Morral JA McComb RD, et al. . The neuropathology of late-onset Friedreich’s ataxia . Cerebellum 2011 ; 10 : 96 – 103 Google Scholar Crossref Search ADS PubMed WorldCat 28 Vankan P. Prevalence gradients of Friedreich’s ataxia and R1b haplotype in Europe co-localize, suggesting a common Palaeolithic origin in the Franco-Cantabrian ice age refuge . J Neurochem 2013 ; 126(suppl 1) : 11 – 20 Google Scholar Crossref Search ADS PubMed WorldCat 2017 American Association of Neuropathologists, Inc. This work is written by US Government employees and is in the public domain in the US. 2017 American Association of Neuropathologists, Inc. This work is written by US Government employees and is in the public domain in the US.
TI - Friedreich Ataxia: Hypoplasia of Spinal Cord and Dorsal Root Ganglia
JF - Journal of Neuropathology & Experimental Neurology
DO - 10.1093/jnen/nlw111
DA - 2017-02-01
UR - https://www.deepdyve.com/lp/oxford-university-press/friedreich-ataxia-hypoplasia-of-spinal-cord-and-dorsal-root-ganglia-0qd891Rzaz
SP - 101
EP - 108
VL - 76
IS - 2
DP - DeepDyve
ER -