An Infant with Pseudohyperkalemia, Hemolysis, and Seizures: Cation-Leaky GLUT1-Deficiency Syndrome due to a SLC2A1 Mutation

Context:

GLUT1 (glucose transporter 1) deficiency syndrome is a well-known presentation in pediatric practice. Very rare mutations not only disable carbohydrate transport but also cause the red cell membrane to be constitutively permeant to monovalent cations, namely sodium and potassium.

Objective:

The aim of this study was to describe the pediatric presentation of a patient with GLUT1 deficiency with such a cation-leaky state.

Subject and Methods:

The infant presented with erratic hyperkalemia, neonatal hyperbilirubinemia, anemia, hepatic dysfunction, and microcephaly. Later, seizures occurred and developmental milestones were delayed. Magnetic resonance imaging and computerized tomography scans of the brain showed multiple abnormalities including periventricular calcification. Visual impairment was present due to the presence of both cataracts and retinal dysfunction.

Results:

Measurements of red cell cation content showed extremely leaky red cells (causing the hemolysis) and temperature-dependent loss of potassium from red cells (explaining the hyperkalemia as pseudohyperkalemia). A trinucleotide deletion in SLC2A1, coding for the deletion of isoleucine 435 or 436 in GLUT1, was identified in the proband.

Conclusion:

This is the fourth pedigree to be described with this most unusual syndrome. The multisystem pathology probably reflects a combination of glucose transport deficiency at the blood-brain barrier (as in typical GLUT1 deficiency) and the deleterious osmotic effects of a cation-leaky membrane protein in the cells where GLUT1 is expressed, notably the red cell. We hope that this detailed description will facilitate rapid diagnosis of this disease entity.

The facilitative, non-sodium-dependent glucose transporter known as “glucose transporter 1” (GLUT1) is the major glucose transporter of the blood-brain barrier. GLUT1 deficiency syndrome (GLUT1DS), caused by defective GLUT1 transport activity at the blood-brain barrier, is a well-known but possibly underdiagnosed pediatric disorder (1, 2) characterized by microcephaly, mental retardation, seizures, and hypoglycorrhachia. The patients are typically heterozygous for mutations in SLC2A1, coding for GLUT1; the clinical condition is attributable to SLC2A1 haploinsufficiency and a gene dose effect at the blood-brain barrier (3). GLUT1 is also expressed in the membrane of the red cell; GLUT1 activity in the red cell can be used as a diagnostic test. The protein is also expressed in lens epithelium (4) and retina (5), but in typical GLUT1DS there is no overt clinical abnormality in the red cell, lens, or retina.

A very small minority of patients with SLC2A1 mutations present with an extended phenotype showing extraneurological pathology. In all of these cases, it has been shown by isotopic tracer studies in the accessible red cell that the membrane is not only defective in glucose transport but also “leaky” to the monovalent cations sodium and potassium. This cation leakiness causes osmotic instability in the erythrocytes and hemolysis (6, 7). Here we give a detailed pediatric description of a child that has a SLC2A1 mutation resulting in deficient glucose transport and red cell cation leakage, leading to pseudohyperkalemia, neonatal hyperbilirubinemia, periventricular calcification on computerized tomography (CT) brain scan, nystagmus, cataracts, and liver dysfunction. There is no direct evidence, but it is presumed that all of the pathology that is additional to the “core” GLUT1DS presentation is attributable the cation-leaky nature of the abnormal protein.

The diagnosis and management of this unusual case caused much clinical difficulty. The case has two lessons: first, it illustrates a novel and surprising presentation of GLUT1DS; and second, the pathology shows how mutations in a facilitative glucose transport can cause abnormal sodium and potassium handling.

Case Report

Background and birth

The infant represented the first pregnancy of an Afro-Caribbean mother and a Caucasian father. She was born at 38.7 wk gestation by emergency cesarean section for fetal distress after a pregnancy complicated by uterine fibroids and polyhydramnios. The birth weight was 3660 g (75th centile), and the head circumference was 32.5 cm (9th centile). There was a paternal family history of adrenoleukodystrophy in two second cousins. The paternal grandmother had been tested and was not a carrier. At 22 h of age, the child presented with severe conjugated hyperbilirubinemia [direct, 411 μmol/liter (24.0 mg/dl); total, 488 μmol/liter (28.5 mg/dl)], abnormal liver function tests (γ-glutamyl transferase, 567 IU/liter), and variable hyperkalemia ranging from normal up to around 10 mmol/liter. Sepsis was suspected and treated for, but cultures were negative. Abdominal ultrasonography showed mild splenomegaly. The toxoplasmosis/rubella/cytomegalovirus/herpes simplex (TORCH) screen was negative; a liver biopsy showed nonspecific hepatitis. The jaundice improved by d 10.

Erratic hyperkalemia

The potassium concentration ([K]) fluctuated from d 1 after birth and was often above 9 mmol/liter. Renal function, acid-base status, and the electrocardiogram were normal. Neither salbutamol, nor insulin-glucose infusions, nor fludrocortisone were effective. Adrenal function, plasma ACTH, plasma renin activity, and aldosterone were all normal (Table 1). It was noted that at least some of the variability in measured plasma potassium was related to the elapsed time between venesection and analysis. Pseudohyperkalemia was suspected and was tested for (see Results and Progress).

Table 1.

Typical hematology and relevant biochemistry, age 4–12 months

Measurement	SI units			US units
	Result	Normal range (for age)	Units	Result	Normal range (for age)	Units
Hematology
Hb	8–12	10.0–13.5	g/dl	8–12	10.0–13.5	g/dl
MCV	96.9	70.0–86.0	fl
MCHC	33.8	31.5–35.5	g/dl
WBC	13.0	6.0–18.0	×109/liter	13.0	6.0–18.0	×103/μl
Platelets	426	150–450	×109/liter	426	150–450	×103/μl
Retiiculocyte count	251	15–110	×109/liter	251	15–110	×103/μl
Haptoglobins	0
Intracellular [Na]	67 cells	5–11	mmol/liter cells	67 cells	5–11	mmol/liter cells
Intracellular [K]	70 cells	88–105	mmol/liter cells	70 cells	88–105	mmol/liter cells
Urea and electrolytes
Na	<140	133–146	mmol/liter	<140	133–146	mEq/liter
K	>5	3.2–6.0	mmol/liter	>5	3.2–6.0	mEq/liter
Urea	2.6	0.7–5.0	mmol/liter	7.3	2.0–14.0	mg/dl
Creatinine	26	13–32	μmol/liter	0.29	0.15–0.36	mg/dl
Ca2+	2.38	2.17–2.44	mmol/liter	4.76	4.34–4.88	mEq/liter
PO4	1.83	1.2–2.1	mmol/liter	5.66	3.71–6.50	mg/dl
Glucose	3.8	3.5–5.5	mmol/liter	68.5	63.1–99	mg/dl
Lactate	1.8	0.7–2.1	mmol/liter	16.2	6.3–18.9	mg/dl
Liver tests
Bilirubin	18	<18	mmol/liter	1.05	<1.05	mg/dl
Alanine transaminase	66	12–47	U/liter	66	12–47	U/liter
Alkaline phosphatase	374	60–330	U/liter	374	60–330	U/liter
γ-glutamyl transferase	112	12–64	U/liter	112	12–64	U/liter
Albumin	43	34–42	g/liter	4.3	3.4–4.2	g/dl
Endocrinology and metabolism
Free T4	12.7	9.0–19.6	pmol/liter	0.97	0.70–1.52	ng/dl
Free T3	7.0	5.1–10.0	pmol/liter	454	331–649	pg/dl
TSH	1.5	<6.0	mU/liter	1.5	<6.0	mU/liter
Aldosterone	990	280–850	nmol/liter	35,4	10.0–30.6	ng/dl
Plasma renin activity	6.9	<10	pmol/ml/h	8.9	<13	ng/ml/h
DHEAS	<0.41	0.5–4.0	μmol/liter	15	18–148	μg/dl
Androstenedione	2.74	3–8	nmol/liter	78	0.90–229	ng/dl
ACTH	16.3	10–50	ng/liter	74	45–227	pg/ml
Synacthen test
Cortisol (zero time)	521	200–600	nmol/liter	18.9	7.2–21.7	μg/dl
Cortisol, +30 min	808	a	nmol/liter	29.3	b	μg/dl
Cortisol, +60 min	822	a	nmol/liter	29.8	b	μg/dl

DHEAS, Dehydroepiandrosterone sulfate; MCHC, mean cell Hg concentration; MCV, mean red cell volume; WBC, white blood count.

^a

Increment of +170 or stimulated level of >550.

^b

Increment of +6.2 or stimulated level of >19.6.

Neurological picture

In the immediate postnatal period, irritability and nystagmus were noted. Milestones were delayed. An electroencephalogram was normal in the first few weeks of life. A brain CT (Fig. 1, A and B) showed periventricular calcification, and an early magnetic resonance imaging scan (Fig. 1, C and D) showed extensive confluent T1w hyperintensity along the lateral margins of both lateral ventricles, extending to the temporal poles bilaterally, which correlated with low T2w signal and some punctate foci of diffusion-weighted imaging signal abnormality. There was thalamic atrophy. The basal ganglia were normal. These appearances were consistent with periventricular calcification, consistent with the CT. There was mild generalized sulcal predominance and thinning of the posterior corpus callosum, in keeping with global loss of white matter volume. There was a normal sulcal/gyral pattern with no evidence of malformation. The cerebellum was of normal volume. There was a slight delay in myelination, with absent myelination within the posterior limb of the internal capsule.

Fig. 1.

Cerebral imaging. A and B, Unenhanced transverse CT showing patchy calcification. C and D, T1-weighted sagittal (C) and transverse (D) magnetic resonance images. Arrows indicate T1w hyperintensity along the lateral margins of both lateral margins (reflecting calcification).

Hearing was normal. At 5.5 months, truncal and four-limb hypertonia were noted. Seizures began at approximately 6 months of age; she was admitted to intensive care in status epilepticus and was paralyzed and ventilated. Blood glucose and calcium concentrations were normal. The seizures, which often happened at night or before feeds, were controlled with phenytoin.

The intracranial calcification and other neurological features suggested a diagnosis of the recessive Aicardi-Goutières syndrome. This recessively inherited leukodystrophic neurodegeneration is caused by mutations in one of a series of immune-related genes (TREX1, RNASEH2A, RNASEH2B, RANSEH2C) (8), but no mutation was found in any of these.

Ocular and visual

Nystagmus was a consistent feature. At age 9 months, nuclear cataracts were noted. The temporal aspects of both discs were considered to be pale, but no other retinal abnormality was noted. At 11 months, a right convergent squint and horizontal jerk nystagmus were documented. Electroretinography and visual-evoked potential studies were consistent with both retinal and postretinal dysfunction.

Hemolysis

Anemia was most severe in the perinatal period [minimum hemoglobin (Hb), 6.8 g/dl], coincident with maximal hyperbilirubinemia. Later, the Hb was typically normal or low-normal. The cells were macrocytic, and reticulocytosis was present. The blood film showed some echinocytes and stomatocytes (Fig. 2A). Splenomegaly was present. As will be described, the hemolysis was later attributed to the cation-leaky state of the erythrocyte membrane.

Fig. 2.

Blood film and pseudohyperkalemia test. A, Wright-Giemsa-stained blood film, showing largely normal red cells with a few echinocytes (e) and occasional stomatocytes (s). B, Pseudohyperkalemia: whole heparinized blood was stored at the temperatures shown. Aliquots were taken at the times indicated and spun, and the supernatant plasma was collected. Plasma [K] was estimated by flame photometry. In the patient's case (upper panel), the [K] was normal at zero time, then rose quickly after venesection, especially at 0 C. In the control case (lower panel), there was only minimal change at any temperature, consistent with clinical experience.

Hepatic dysfunction

Hyperbilirubinemia was always present and was attributable to hemolysis. However, the alanine transaminase and γ-glutamyl transferase were always at least slightly abnormal. At 4 wk, the liver biopsy showed nonspecific hepatitis.

Method: DNA sequencing analysis

Genomic DNA was isolated from blood samples. The coding regions and splice sites of exons 1 to 10 of SLC2A1 from the child and exon 10 of SLC2A1 from the parents were amplified by PCR using exon-specific primers. The DNA was sequenced as previously described (7).

Results and Progress

In a test for pseudohyperkalemia, whole heparinized blood of the patient and an adult control was stored at 37, 20, and 0 C for up to 6 h (9). Aliquots were taken from the samples and centrifuged at 0, 1, 2, 4, and 6 h after venipuncture, and the [K] was measured. The results are shown in Fig. 2B. In the control sample, there was no major change in plasma [K] at any temperature [as is normal (9)]. By contrast, in the infant's sample, the plasma [K] rose at all three temperatures. The rise was most striking at 0 C, confirming that this blood, when stored, shows an abnormal rise in plasma [K] due to leakage of potassium from red cells. The very rapid rise at 0 C is consistent with the cryohydrocytosis phenotype of stomatocytosis (9). Measurements of intracellular sodium and potassium revealed a gross abnormality in fresh red cells: the intracellular sodium was very high, and the potassium was low (Table 1). These data showed that the cells were very leaky to these cations. Intracellular sodium concentration ([Na]) and [K] in the parents' red cells were normal (data not shown). These data showed that, first, the hyperkalemia was due to a temperature-dependent artifact, and second, that the infant had markedly abnormal red cells with a very dramatic abnormality in intracellular [Na] and [K], even in fresh cells, consistent both with the hemolysis and with a diagnosis of cryohydrocytosis, a member of the cation-leaky hereditary stomatocytosis group (10).

Because we had previously shown that two adult patients with neurological retardation and abnormal cation-leaky red cells (11) had mutations in SLC2A1 encoding for GLUT1 (7), we sequenced SLC2A1. The infant was heterozygous for an ATC trinucleotide deletion, coding for isoleucine 435 or 436. The parents were normal. This is the same mutation that was found in a previous adult patient labeled sdCHC(B) (7).

The corollary of this finding was that at least part of the neurological presentation should be due to deficiency of GLUT1-mediated transport of glucose across the blood-brain barrier. A fasting lumbar puncture was performed. The cerebrospinal fluid glucose concentration was 2.0 mmol/liter (36.0 mg/dl) compared with a simultaneous plasma glucose concentration of 3.8 mmol/liter (68.5 mg/dl), confirming the presence of hypoglycorrhachia, consistent with deficient transport of glucose across the blood-brain barrier.

Western blots showed that GLUT1 was expressed at normal levels in the red cell membrane of the affected child, suggesting that the mutant protein is expressed and stably inserted into the membrane (Fig. 3B). The cells showed reduced expression of the 32-kDa protein stomatin (Fig. 3C), a “raft” protein of poorly understood function that is mistrafficked in cation-leaky red cells (12). “Lutheran” is an Ig-like protein of the red cell membrane that acts as a receptor for laminin (13). As in the previous cases (7), the expression of Lutheran was increased in the abnormal cells (Fig. 3D). There was no abnormality in the membranes of the hematologically normal parents (data not shown).

Fig. 3.

Western blots. Red cell membranes were prepared by hypotonic lysis (18), separated on 10% Laemmli gels, and immunoblotted using the indicated antibodies as described (7). C lanes, Control; P lanes, proband. A, Coomassie stain. The absence of a Coomassie staining band at 32 kDa represents the deficiency of stomatin, confirmed in blot C. B, Anti-GLUT1 antibody. The GLUT1 protein is heavily glycosylated and is seen as a broad 50- to 100-kDa band. In the sdCHC patient, the amount of GLUT1 was not reduced, suggesting that the mutant glut1 is expressed normally. C, Anti-stomatin antibody. The protein is reduced in expression but not entirely absent. D, Anti-Lutheran antibody. The Lutheran protein was increased in the sdCHC sample. An actin (or protein 4.2) loading control is shown beneath the blots.

Thus, an unusual form of GLUT1 deficiency was diagnosed. The infant was started on a ketogenic diet. The seizure frequency decreased and, in the opinion of both parents and clinicians, the infant became less irritable. The hyperkalemia was understood to be red cell-based, temperature-dependent pseudohyperkalemia. It was ensured that the elapsed time between venipuncture and analysis was minimized, and measured potassium levels were thereafter always normal.

Discussion

This case confirms the association between a hematoneurological syndrome and mutations in SLC2A1 and, for the first time, illustrates the pediatric presentation in detail. As described for the vast majority of SLC2A1 mutations, the patient is heterozygous, and the mutation is de novo (1–3). As shown in the previous adult cation-leaky case (7), the altered conformation of the GLUT1 protein has two qualities: first, there is loss of glucose transport function; and second, there is occurrence of a nonspecific “leak” to the monovalent cations sodium and potassium, which can be seen and measured in red cells.

The dual nature of the mutant GLUT1 protein can account for the complex clinical picture in this child. First, there is the picture of GLUT1DS, caused by the haploinsufficiency of glucose transport at the blood-brain barrier (1, 3), which could account for microcephaly, developmental delay, and seizures and is confirmed by hypoglycorrhachia and the positive therapeutic response to the ketogenic diet. Second, and not seen in the simple loss of function GLUT1DS, there is hemolytic anemia due to a cation leak, with associated pseudohyperkalemia attributable to the bizarre temperature dependence of the cation leak. In addition to the hemolysis, this child demonstrates markedly abnormal brain imaging with periventricular calcification, leading to a suspected diagnosis of Aicardi-Goutières syndrome (8). There was also visual impairment, with probable retinal dysfunction and cataracts, and hepatic dysfunction. In our two adult cases, cataracts and hepatosplenomegaly were noted, but no brain imaging was available (11).

Although the hemolysis was certainly attributable to the cation leak, cation movements are difficult to assess in the other tissues and cell types. GLUT1 is expressed in lens epithelium (4) and retina (5). It is possible that the periventricular calcification, which is probably a form of dystrophic calcification (14), is due to cellular injury at the blood-brain barrier; the hepatic dysfunction could be due to a cation leak in cells within the liver. This is speculation because measurements of cation content and flux are not possible in these human tissues.

The hematological findings in this child are similar to those observed in our previous cases (11), with the exception of the red cell morphology. The cells are very leaky to sodium and potassium at 37 C, as evidenced by the very abnormal intracellular [Na] and [K] in the fresh cells. On storage at 0 C, the cells very quickly lose potassium, consistent with a cryohydrocytosis phenotype, in which the cells are leakiest at 0 C (9, 11). The cells also lose potassium when stored in heparin at 37 C; this is probably due to energy deprivation, the very active cells quickly using up glucose in the surrounding plasma. There is minimal glucose utilization at 0 C, so this point cannot explain the loss of potassium at low temperatures.

Pseudohyperkalemia (an artifactual rise in apparent plasma potassium due to in vitro loss of potassium from blood cells) is a well-recognized aspect of leaky cell conditions of the hereditary stomatocytosis group (10). It was very marked in this case and caused much clinical consternation, especially because the infant had seizures. Pseudohyperkalemia can be recognized by the random nature of the readings, by the lack of background pathology (renal or adrenal failure, drug treatment, diabetes, acidosis), and by the recognition that those samples that show the highest readings have been stored the longest before analysis.

Phenotypic variation is a feature of GLUT1DS (1), and this quality applies to these cation-leaky cases. Two individual patients and one pedigree have been described with cation-leaky GLUT1DS. In the pedigree, which showed the deletion Q282-S285 in GLUT1, the neurology was less severe, manifesting as paroxysmal exercise-induced dyskinesia (6). As in the present case, the red cell morphology was echinocytic. In the male adult case previously described by us, labeled sdCHC(A) (7), who showed the substitution gly286asp, the neurology was similarly severe to the present case, and the red cell morphology was stomatocytic. In case sdCHC(B) (7), where the mutation was the same as in the present case, the neurology was likewise severe, but the morphology was not described. The explanation for this difference in red cell morphology is unclear. Variations in modifying genes are possible, but it is already known that there is marked phenotypic variability in the GLUT1DS phenotype dependent on the exact SLC2A1 variation.

Note that this condition is different from neuroacanthocytosis, which also shows spiculated red cells with a neurological phenotype. In neuroacanthocytosis, the onset of neurological symptoms occurs later in life, and the conditions are caused by mutations in VPS13A, JPH3, XK, or PANK2 (15).

As in our previous cases (11), the stomatin protein was deficient from the membrane. Mistrafficking of stomatin during erythropoiesis is seen in these very leaky GLUT1 mutants, in the rhesus-associated glycoprotein mutants (16), and in dog red cells with a genetic polymorphism that down-regulates the NaK pump (17). The mechanism and purpose behind this mistrafficking is not understood; for the present case, stomatin can be considered to be a biomarker for cells that are high in intracellular sodium and low in potassium.

We hope that this report will facilitate diagnosis in similar cases.

Acknowledgments

We thank the parents for permission to publish this report. Many clinicians contributed to the care of the infant, including Dr. C. D. Notaney (Wembley), Prof. Anil Dhawan (Kings College Hospital), Dr. Courtney Wusthoff (Hammersmith Hospital), Dr. Sepali Wijesinghe (Brent), and Dr. Alki Liasis (Great Ormond Street Hospital). We thank Dr. Bari Levinson (San Francisco) for invaluable advice. We thank Advocacy for Neuroacanthocytosis for generous support.

This work was supported by Advocacy for Neuroacanthocytosis (to G.W.S.) and by the United Kingdom National Health Service R&D Directorate (to W.M.B., A.L.A., J.F.F., and L.J.B.).

Author Contributions: A.L.A., W.M.B., and J.F.F., protein gel analysis and gene sequencing; B.J., C.O., and S.G., key clinical contributions; E.F.G., M.D., G.W.S., and L.J.B., authorship.

Disclosure Summary: No author has any conflict of interest to report.

Abbreviations

 
• CT

  Computerized tomography

 
• GLUT1

  glucose transporter 1

 
• GLUT1DS

  GLUT1 deficiency syndrome

 
• Hb

  hemoglobin

 
• [K]

  potassium concentration

 
• [Na]

  sodium concentration.