Iron Dysregulation in Restless Legs Syndrome

Chapter 10 Iron Dysregulation in Restless Legs Syndrome

In the 1940s and 1950s, while Ekbom^1,2 was proposing a vascular model for restless legs syndrome (RLS), Nordlander³ was establishing a role for iron in RLS. He noted that his patients with RLS had a high prevalence of iron deficiency and that their symptoms improved on treatment of the iron deficiency. In 1953, Nordlander formally proposed “that restless legs might … be a sign of iron deficiency.”³ Ekbom acknowledged Nordlander’s finding and reported a 24.6% prevalence of iron deficiency among his patient population.⁴ O’Keefe and colleagues⁵ also reported a high prevalence of iron deficiency among their patients with RLS. In addition, there is a high prevalence of RLS in populations with iron deficiency (20% to 30%),^4,6 in populations associated with a high prevalence of iron deficiency like pregnancy^7–9 and blood donors (14% to 26%)^10,11 and in populations with altered iron metabolism like occurs in dialysis (14% to 68%).^12–14 Although the elderly may have low iron stores due to aging or chronic disease and therefore suffer RLS,⁵ in the pediatric population, where RLS may be associated with attention-deficit/hyperactivity disorder (ADHD),^15–17 low iron stores measured by serum ferritin levels have been associated with ADHD,¹⁷ RLS,¹⁶ and comorbid ADHD and RLS.^16–18

Two independent studies demonstrated a strong inverse correlation between serum ferritin levels and RLS severity: as serum ferritin declined, symptom severity increased.^5,19 One of the studies also showed a strong correlation between serum ferritin level and sleep efficiency: as ferritin declined, sleep efficiency also declined.¹⁹ Despite the findings of these two studies there are limitations to their interpretation. Both evaluated relatively small populations. O’Keefe and associates⁵ studied an elderly population where the majority of the patients had low-normal or abnormal ferritin levels with a median ferritin level for the group of 33 µg/L (range of 6 to 124 µg/L). Although the patient population studies by Sun and coworkers¹⁹ had a much broader range of ferritin levels (5 to 229 µg/L), the mean ferritin was in the low-normal range (60 µg/L). On further analysis of these data,¹⁹ patients with low ferritin levels (<50 µg/L) had greater disease severity by several different outcomes measures compared with those with high ferritin levels (>50 µg/L). This finding and others²⁰ suggest that the relation between RLS symptoms and a broad range of body iron storage states may be poor. The relation is strongest when the patients have low or deficient body iron stores, suggesting that iron deficiency per se may be more important than general iron status.

Nordlander³ successively treated RLS symptoms with intravenous iron. The surprising fact is that the majority of these patients had normal blood iron levels. The study implicates a causal link between iron and RLS symptoms and supports his initial premise that “there can exist an iron deficiency in the tissues in spite of normal serum iron.”³ More recent studies have considered the possibility that the brain per se could have low iron stores despite the presence of normal blood or systemic levels of iron. Using magnetic resonance imaging techniques to quantify iron concentrations in 11 brain regions, a first study of RLS patients found lower iron concentrations in their substantia nigra than in age-matched control subjects.^21,22 In addition, lower iron concentrations in the substantia nigra were significantly correlated with increasing RLS symptom severity. A second, expanded study confirmed this finding but primarily for early-onset RLS.²³ Several studies using transcranial Doppler have confirmed the presence of low levels of iron in the substantia nigra of RLS patients.^24–26 The distinction between regulation of body and brain iron levels is made most strikingly by the finding that patients with hemochromatosis—who retain iron and have high ferritin levels—may have reduced brainstem iron.²⁷

Another study examined cerebrospinal fluid (CSF) ferritin and transferrin as the indicator of brain iron status²⁸ (Figs. 10-1, 10-2, and 10-3 ). Subjects were chosen who had normal blood levels of both of these factors. Despite having normal systemic iron stores that were comparable with those of control subjects, patients with RLS had reduced CSF ferritin and increased CSF transferrin, findings that suggest a relative iron-insufficient state within the central nervous system (CNS). The correlation between serum and CSF ferritin was different for control subjects and RLS patients. The slope of the curve in the RLS group was lower than that found in the control group. This would suggest that CNS iron regulation might be somewhat independent of the systemic iron status in RLS patients. In additional studies using immunoblot techniques and controlling for overall protein quantity, decreased H and L subunits were found in early-, but not late-onset RLS.²⁹ Similarly, depressed levels of prohepcidin, a precursor of the iron regulatory protein hepcidin, were also found in early-onset patients.³⁰

FIGURE 10-1. Serum concentrations for iron, ferritin, and transferrin in patients with restless legs syndrome (RLS) and in control subjects (Normal). The mean for the individual values is indicated by the blue bar.

Reprinted with permission from the American Academy of Neurology. Source: Earley CJ, Connor JR, Beard JL, et al. Abnormalities in CSF concentrations of ferritin and transferrin in restless legs syndrome. Neurology 2000;54:1698-1700.

FIGURE 10-2. CSF concentrations for iron, ferritin, and transferrin in patients with restless legs syndrome (RLS) and control subjects (Normal). The mean for the individual values is indicated by the blue bar.

Reprinted with permission from the American Academy of Neurology. Source: Earley CJ, Connor JR, Beard JL, et al. Abnormalities in CSF concentrations of ferritin and transferrin in restless legs syndrome. Neurology 2000;54:1698-1700.

FIGURE 10-3. A linear regression plot of the relation between serum ferritin and cerebrospinal fluid ferritin for patients with restless legs syndrome (RLS; •) and control subjects ().

Reprinted with permission from the American Academy of Neurology. Source: Earley CJ, Connor JR, Beard JL, et al. Abnormalities in CSF concentrations of ferritin and transferrin in restless legs syndrome. Neurology 2000;54:1698-1700.

The most conclusive data on the relation of brain iron metabolism to RLS are from brain autopsy analyses³¹ (Figs. 10-4 and 10-5). General histological assessment of brains from seven RLS patients and five control subjects showed no obvious pathology, cell loss, gliosis, or signs of Parkinson’s or Alzheimer’s disease. Assessments of iron and iron-regulator proteins were specifically evaluated in the substantia nigra because of its high iron and dopaminergic concentrations and because of the previous magnetic resonance imaging findings indicating low iron levels in this area in RLS. In tissues from patients with RLS, iron and H-ferritin were decreased, transferrin was increased, and L-ferritin was the same as for control subjects. However, L-ferritin appeared more concentrated in glial cells in RLS tissue, whereas L-ferritin was predominantly found in oligodendrocytes of control tissues.

FIGURE 10-4. Representative micrographs from the human substantia nigra comparing control brains (left) with restless legs syndrome (RLS) brains (right). (A, B) Tyrosine hydroxylase. The number of substantia nigra neuromelanin cells is comparable in both the control (A) and RLS (B) brains. The neuromelanin is the brown pigment in the cells. The cell bodies and their processes are immunolabeled for tyrosine hydroxylase (chromagen for the reaction product is blue) with no visibly detectable difference between control and RLS brains. The arrows point to typical cells as examples. Bar = 12.5 μm. (C, D) Iron staining. (C) A modified Perls reaction for iron staining is used (see text), and in this figure the neuromelanin is brown and the iron reaction product is blue. In the control brain, the neuromelanin-containing cells are clearly present and contain no blue reaction product for iron (e.g., the cell at the white arrowhead containing brown neuromelanin pigment). The iron-positive cells (e.g., at arrow) are oligodendrocytes. (D) In the RLS brain, the neuromelanin cells do not stain for iron and there are relatively few iron-positive cells. Occasionally, cells that stain for the Perls reaction are present in the RLS brain. An example of this cell type is indicated in the figure (arrow) and has processes, unlike those iron-positive cells seen in the control brains. Bar = 16 μm. (E, F) H-ferritin. In these figures, the H-ferritin reaction product is blue. The neuromelanin pigment is brown. (E) In the control brain there is blue reaction product in the neuromelanin cells (e.g., at white arrowhead) indicating the presence of H-ferritin. In addition, numerous blue staining processes and cells are present in the parenchyma. Occasionally, an H-ferritin–positive oligodendrocyte is visible (arrow). (F) In the RLS brain section, there is no detectable H-ferritin in the neuromelanin-containing neurons or the processes in the parenchyma. A few positive glial cells are visible (e.g., at arrows). This observation is consistent with an absence of stored iron in the neurons in this region of the RLS brain. Bar = 16 μm. (G, H) L-ferritin. In this figure, L-ferritin immunoreaction product is blue and the neuromelanin is brown. (G) L-ferritin is present in the control brains in small round cells (arrows) but rarely in cell processes. Most of the neuromelanin-containing cells do not stain for L-ferritin. (H) In the RLS brain, L-ferritin–positive cells are also clearly present but are morphologically distinct from the majority of cells in the control brains. Almost all of the L-ferritin–positive cells in the RLS brain are ramified (arrows) and have the morphological appearance of astrocytes and microglia. L-ferritin–positive neuromelanin cells are rare. Bar = 8 μm.

Reprinted with permission from the American Academy of Neurology. Source: Connor JR, Boyer PJ, Menzies SL, et al. Neuropathological examination suggests impaired brain iron acquisition in restless legs syndrome. Neurology 2003;61:304-309.

FIGURE 10-5. Micrographs from the human substantia nigra of proteins involved in iron transport. Micrographs from control brains are on the left and micrographs from restless legs syndrome (RLS) brains are on the right. (A, B) Metal transport protein 1 (MTP1). The control brain (A) shows the presence of MTP1 (blue reaction product) in the neuromelanin cells. The MTP1 reaction product is confined mostly to the soma (e.g., at arrows), but a few primary processes are visible. The melanin pigment appears brown in the micrograph. There is a considerable amount of process staining in the neuropil. (B) In the RLS brain, the neuromelanin pigment is brown. Most of the neuromelanin cells are devoid of reaction product (e.g., cell at arrow) but immunostained processes in the neuropil are prominent. These immunostained processes have a similar appearance to those in the control brains. Bar = 8 μm. (C, D) Divalent metal transporter 1 (DMT1). In C, the staining in the control brains (blue reaction product) shows a strong reaction in the brown neuromelanin-containing cells extending into the primary process (e.g., cell at arrow), whereas the staining in the RLS brains (D) is rare with only an occasional small cell (e.g., at the arrow) present. Bar = 30 μm. (E, F) Transferrin receptor. (E) Transferrin receptor immunoreaction product (blue) is present on most of the brown neuromelanin containing cells in the control brains. The immunoreaction product is found on the soma and extends into a primary process in all cases (e.g., cell at arrow). (F) In the RLS brains, the immunoreaction product for the Tf receptor is minimal and immunostained cell processes are rare on the brown neuromelanin cells (e.g., cell at arrow). Bar = 16 μm. (G, H) Transferrin. (G) Transferrin (blue reaction product) is visible in the neuromelanin cells (e.g., at arrow) in the control brain and in oligodendrocytes, some of which are near the neuromelanin-containing neurons (e.g., arrowhead). There are also processes that are Tf positive in the neuropil. (H) By comparison, the RLS brain has considerably more Tf reaction product in the neuromelanin cells (e.g., at arrow) and in the processes of these cells than the control brain. The Tf-positive processes in the neuropil are also much more striking in the RLS brain than in the control brain. Occasional small, round cells are present (e.g., arrowhead) as in the control brain. Bar = 8 μm.

Reprinted with permission from American Academy of Neurology. Source: Connor JR, Boyer PJ, Menzies SL, et al. Neuropathological examination suggests impaired brain iron acquisition in restless legs syndrome. Neurology 2003;61:304-309.

A surprising finding was the normal or low-normal levels of transferrin receptor (TfR) protein in RLS tissues. In response to the low tissue iron, TfR protein should have increased to capture more iron. More quantitative analyses performed in neuromelanin cells isolated from the substantia nigra using laser-capture, microdissection techniques showed findings identical to those found with the immunohistochemical technique, including decreased TfR protein³² (Fig. 10-6).

FIGURE 10-6. Quantitative immunoblot assays are performed to determine the relative amounts of the iron-management proteins in neuromelanin-containing neurons isolated by laser capture microdissection. Heavy chain ferritin (A), divalent metal transporter 1 (C), ferroportin (D), and transferrin receptor (E) are decreased in restless legs syndrome (RLS) neuromelanin cells. The relative amount of light chain ferritin (B) is similar between the two groups. Transferrin (F) is increased in RLS neuromelanin cells. The triangles represent the mean value for each individual obtained by performing the assays in triplicate. The blue square is the mean of each group. The error bars represent the SD for each individual from the triplicate assays and the SD for the group. *p < 0.05; ***p < 0.001.

Reprinted with permission of the American Academy of Neurology. Source: Connor JR, Wang XS, Patton SM, et al. Decreased transferrin receptor expression by neuromelanin cells in restless legs syndrome. Neurology 2004;62:1563-1567.

As discussed in Chapter 9, ferritin and TfR synthesis are controlled by one of two iron-responsive proteins (IRP1 and IRP2). Under conditions of iron deficiency, there is increased synthesis of IRP2 and increased conversion of aconitase to IRP1. Increased levels of IRP then result in decreased ferritin and increased TfR. Therefore, to further understand the role of iron-regulator proteins in RLS, the protein levels of both IRPs and aconitase were analyzed in neuromelanin cells that had been isolated from the substantia nigra using laser-capture, microdissection techniques³² (Figs. 10-7 and 10-8).

FIGURE 10-7. The relative concentrations of iron regulatory protein 1 (IRP1) (A) and IRP2 (B) in neuromelanin-containing neurons were determined by immunoblot assay. The triangles represent the mean of each individual, and the squares represent the mean of the group. The error bars represent the SD for each of the individual samples and for the entire group. Because IRP1 has aconitase activity when it is not in the mRNA binding form, an aconitase assay was performed (C). The data are plotted as described in the legend for A and B. **p < 0.01; ***p < 0.001.

Reprinted with permission of the American Academy of Neurology. Source: Connor JR, Wang XS, Patton SM, et al. Decreased transferrin receptor expression by neuromelanin cells in restless legs syndrome. Neurology 2004;62:1563-1567.

FIGURE 10-8. Iron regulatory protein (IRP)-IRE binding activity was determined in the cell homogenates from isolated neuromelanin cells. (A) The amount of IRP available in the cell homogenates to an IRE probe was similar between the two groups. (B) The amount of IRP available to bind to an IRE probe is determined by incubating the cell homogenates with β-mercaptoethanol. The available IRP is significantly decreased in the cells from the patients with restless legs syndrome (RLS) compared with control. (C) To determine the relative amounts of the two IRPs that are binding to the IRE probe, binding assays were performed in the presence of IRP2 antisera. This figure shows the amount of binding activity that migrated at a slower rate in the gel-shift assay as a result of the addition of IRP2 antibody is similar between the two groups. (D) The amount of IRP activity in the gel-shift assay that was unaffected (unshifted) by the presence of the IRP2 antisera is decreased in RLS compared with control. These data indicate that less IRP1 is present in the bound form in the cell homogenates from the RLS brains compared with control. *p < 0.05.

Reprinted with permission of the American Academy of Neurology. Source: Connor JR, Wang XS, Patton SM, et al. Decreased transferrin receptor expression by neuromelanin cells in restless legs syndrome. Neurology 2004;62:1563-1567.

Neuromelanin cells from RLS tissue had significantly lower IRP1 and elevated IRP2 levels compared with control tissue. The aconitase concentration and its activity were also diminished in RLS tissue. IRP1 seems to be more important for the regulation of TfR synthesis than IRP2.³³ This does not appear to be true for IRP control of H-ferritin synthesis. Therefore, lower IRP1 leads to a diminished TfR synthesis, whereas increased IRP2 resulted in the expected decreased H-ferritin synthesis. The basis for the inadequate response or quantitatively diminished IRP1-aconitase remains unknown at this time. The findings would implicate disruption of iron regulation at the level of IRP1 or higher.

A major consideration is how any brain deficiency in RLS causes the emergence of the syndrome. Several lines of evidence suggest how this might occur through alteration of the dopamine system. First, iron can lead to dysfunction in regulation of dopamine system activity. CSF studies of dopamine metabolites show a greater difference between evening and morning values in RLS patients than in control subjects,³⁴ suggesting an enhanced circadian variation of RLS and alterations in metabolism and perhaps increased dopamine synthesis or turnover. Second, a key protein for synapse formation and maintenance, Thy-1, is reduced in RLS dopamine neurons in the substantia nigra.³³ Moreover, in animal models, iron deprivation can induce increased circadian variability of dopamine.³⁵

In summary, iron appears to play a key role in RLS.³⁶ There are probably several different triggers by which iron dysregulation may occur. Some of the mechanisms for control of brain iron have been explored in mice and rats and are discussed in Chapter 9. When systemic iron deficiency is induced in different strains of mice, the brain response is quite variable, with some strains showing no change in brain iron whereas others showed decreases in iron.³⁷ In susceptible strains, iron deficiency can produce circadian changes in activity that mimic RLS in showing increased activity at the transition to the rodent rest period.³⁸ A similar finding has been noted in animal models with lesioning of the A11 dopamine system that projects to the spinal cord:^39,40 iron deprivation causes greater changes in dopamine receptors and enhanced activity.⁴¹ Similarly, in humans, a primary systemic iron deficiency could lead to secondary brain iron insufficiency in a subgroup of the population and thus lead to RLS symptoms. But what about the larger RLS population for whom systemic iron deficiency does not appear to play a role?

When ventral midbrain iron is analyzed across 30 mice strains that were developed from the BXD recombinant inbred mice colony, one can see a three-fold difference in the highest to lowest ventral midbrain iron concentration which had no relation to the systemic levels of iron.⁴² This would indicate the presence of a complex genetic regulation of brain iron concentrations at the level of the blood-brain barrier or within the cells of the brain that can selectively produce low iron levels without any significant changes in systemic iron concentrations. The parallel to the human condition is important, because it shows the complex genetic nature of brain iron regulation that may exist in humans and account for the changes seen in the RLS brain, despite normal or even increased²⁷ systemic levels of iron.