Cerebral and portal vein thrombosis, macrocephaly and atypical absence seizures in Glycosylphosphatidyl inositol deficiency due to a PIGM promoter mutation

Abstract

Defects of the glycosylphosphatidylinositol (GPI) biosynthesis pathway constitute an emerging subgroup of congenital disorders of glycosylation with heterogeneous phenotypes. A mutation in the promoter of PIGM, resulting in a syndrome with portal vein thrombosis and persistent absence seizures, was previously described in three patients. We now report four additional patients in two unrelated families, with further clinical, biochemical and molecular delineation of this unique entity. We also describe the first prenatal diagnosis of PIGM deficiency, allowing characterization of the natural history of the disease from birth. The patients described herein expand the phenotypic spectrum of PIGM deficiency to include macrocephaly and infantile-onset cerebrovascular thrombotic events. Finally, we offer insights regarding targeted treatment of this rare disorder with sodium phenylbutyrate.

Keywords: Absence seizures; Congenital disorders of glycosylation; GPI; Glycosylphosphatidylinositol deficiency; PIGM; Portal vein thrombosis.

Figure 1 –

(A) Individuals with PIGM mutations. (1) Siblings from Family A (from left to right: unaffected sister A-IV:2, affected sister A-IV:4 and affected brother A-IV:3). (2) Patient A-IV:4 at age 10 years. (3) Patient B-IV:1 at the age of 2 years. Note the prominent superficial veins in the faces of all affected individuals, as well as a relatively large head circumference. (B) Pedigrees of Families A and B, with the c.−270C>G variant status of all individuals tested. Full symbols designate affected individuals. Partially full symbols in Family B designate individuals reported to have macrocephaly, without additional findings. +/+, +/−, and −/− signs designate homozygosity, heterozygosity and wild type status for the c. −270C>G variant in PIGM.

Figure 2 –

Brain MRI images of patients with PIGM deficiency, demonstrating cerebral infarcts in 3 of 4 patients. (A) T2 FLAIR axial section in patient A-IV:4 demonstrating hyperintense signal and mild atrophy of right parietal cortex delineating an old cortical infarct. (B) T2 axial section in patient A-IV:4 demonstrating small right frontal old hemorrhage. (C) T1 flair axial section in patient A-IV:9 demonstrating a large right cerebral infarct in MCA distribution. (D) T1-weighted axial section in patient A-IV:3 demonstrating general cortical atrophy and an old left occipital infarct. (E) T2 FLAIR axial section in patient B:IV:1 demonstrating mild cortical atrophy and few small frontal lesions (white arrows) with hyperintense signal suggestive of micro-ischemic events.

Figure 3 –

Electoencephalographic (EEG) recordings of two patients with PIGM deficiency, demonstrating epileptiform activity. (A) EEG trace from patient A-IV:4 demonstrating 2.5–3 Hz generalized spike wave bursts. (B) EEG trace from patient A-IV:4 demonstrating encephalopathic background of spike and slow wave activity with frontal predominance. (C) EEG trace from patient A-IV:3 demonstrating encephalopathic background of spike and slow wave activity with right predominance. All traces are shown at voltage amplitude of 15μV/mm and timescale of 30 mm/sec.

Figure 4 –

Relative PIGM mRNA expression levels as determined by qPCR in affected individuals and healthy controls blood (B) and fibroblast cell lines (F). PIGM gene expression levels in blood samples from controls were considered to be 100% and all other samples were compared to it. Asterisk indicates statistically significant differences of relative PIGM mRNA levels (P ≤ 0.05) between controls and patients according to t test. Vertical lines represent standard error values.

Figure 5 –

Expression of GPI and GPI-linked proteins on blood cells during sodium phenylbutyrate therapy. Sodium phenylbutyrate was started at a dose of 10 mg/kg three times daily (TID) (30 mg per kilogram of body weight, for 24 h), increased to 20mg/kg TID (60 mg/kg/24h) after two weeks and finally increased to 30 mg/kg TID (90 mg/kg/24h) (first arrow indicates the date of treatment initiation and the second and third arrows indicate dosage elevation). For patients A-IV:4 and A-IV:9 treatment was started at the final dose of 90 mg/kg/d. Blood from each patient was taken at least once a month and analyzed for GPI and GPI-linked antigens on granulocytes, monocytes and erythrocytes (i.e.CD59, CD24, CD14 and FLAER) via FACS. For patient A-IV-4, a gradual elevation in GPI and GPI-linked proteins on granulocyte was observed upon treatment (upper graph). For patient A-IV-9, an initial decrease followed by a mild increase in GPI and GPI-linked proteins was observed. For patient B-IV-1, there was an initial increase in the percentage of cells stained positive for GPI and GPI-linked proteins upon treatment initiation followed by return to baseline within 5 month of treatment (lower graph). For all three patients no significant change was observed in erythrocytes or monocytes (red and green graphs, respectively).

Figure 6 –

Decreased expression of GPI-anchored proteins CD59 and CD87 on patient fibroblasts. Patient cells show a marked reduction in membrane bound CD59 (green, top row) compared to control fibroblasts. Although overall expression of CD87 (green, bottom row) is low in both control and patient fibroblasts, expression of this GPI-anchored protein on the membranes of patient fibroblasts is reduced compared to controls. Nuclei were co-stained with Hoescht 33342 (blue).