Retinal Vasculopathy With Cerebral Leukodystrophy: Clinicopathologic Features of an Autopsied Patient With a Heterozygous TREX 1 Mutation

Abstract

INTRODUCTION

Retinal vasculopathy with cerebral leukodystrophy (RVCL) is an adult-onset autosomal-dominant small vessel disease affecting microvessels in the cerebrum, retina, kidney, and other multiple organs (1). RVCL is now recognized to encompass several clinical syndromes, including (i) hereditary vascular retinopathy (ii) cerebroretinal vasculopathy (iii) hereditary endotheliopathy, retinopathy, nephropathy and stroke (HERNS) (4), and (iv) hereditary systemic angiopathy (HSA) (5). RVCL is caused by heterozygous C-terminal frameshift mutations in TREX1 encoding 3′–5′ DNA exonuclease (6). The TREX1 protein, part of the SET complex, is ubiquitously expressed in mammalian cells, where it mediates DNA damage—preferentially single-stranded DNA—in association with granzyme A (7). In normal cells, TREX1 resides within the cytoplasm and the C-terminal domain is required for cytoplasmic localization but can translocate into the nucleus in response to oxidative stress in order to repair DNA damage (8). The localization and kinetics of TREX1 are important for maintaining the physiological functions of the cell.

In vitro studies of RVCL using transient transfection in cultured cells have shown that the mutated TREX1 protein lacking the C-terminal domain for anchoring to the endoplasmic reticulum retains enzymatic activity and becomes diffusely distributed in the cytoplasm and nucleus (6). Considering its physiological function, the mislocalization of TREX1 protein is thought to be intrinsically linked to the development of RVCL pathology. However, little is known about its endogenous expression in the affected human brain and other organs. Here, we describe in detail the clinicopathologic features of an autopsied patient with RVCL and discuss the relationship between aberrant localization of the mutated protein and the pathogenesis.

CASE PRESENTATION

A 26-year-old Japanese man presented with proteinuria. His mother had been blind and died of a brain tumor in her thirties. His daughter had asymptomatic proteinuria and had undergone renal biopsy; details of the renal histopathology have been reported previously (9). Renal biopsy of the present patient revealed glomerulonephropathy with a background histology resembling membranoproliferative glomerulonephritis, which was similar to that found in his daughter. At the age of 30 years he noted visual disturbance. Fluorescein angiography revealed bilateral retinopathy. Pan-retinal laser photocoagulation was applied, but bilateral vitreous hemorrhage occurred and he lost the sight in his left eye. Three months after retinal therapy, he was admitted to hospital due to sudden onset of right hemiplegia. Brain CT showed a diffuse low-density area in the left frontal and temporal white matter. Calcification was not evident. Enhanced CT revealed a ring-enhancing solid tumor-like mass with extensive cerebral edema in the frontal deep white matter on the left side. The cerebrospinal fluid cell count and protein level were normal. Laboratory findings indicated end-stage renal failure manifested by elevation of the serum creatinine level to 8.1 mg/dL, urea nitrogen 90 mg/dL, and proteinuria (5 g/24 hours). Serum IgG, IgA, IgM, C3, and C4 levels were within normal limits. Rheumatoid factor and antinuclear antibodies were not detected. The patient underwent a biopsy of the left cerebral hemispheric lesion. The histopathological features revealed ischemic changes and slight perivascular cuffing of inflammatory cells. There were no tumor cells. Glucocorticosteroid was then administered to reduce the cerebral edema and inflammation. This resulted in dramatic resolution of the hemiparesis and a clinically stable period lasting a few months. However, a generalized epileptic seizure occurred 2 months after the brain biopsy. CT scan of the brain demonstrated expansion of the frontal low-density area and focal calcifications in the left frontal white matter. At the age of 37 years, the patient died of thrombocytopenia and sepsis 11 years after initial admission to hospital. There had been no history of oral, genital, or skin lesions, sicca symptoms, or musculoskeletal, respiratory symptoms. A general autopsy was performed, at which time the brain weighed 1245 g.

MATERIALS AND METHODS

The present study was approved by the Ethics Committee of Niigata University. Informed consent for autopsy, collection of samples and their subsequent use for genetic analysis and other research purposes was obtained from the patient’s family.

The brain and spinal cord were fixed with 20% buffered formalin, and multiple tissue blocks were embedded in paraffin. Histological examination was performed on 4-μm-thick sections using several stains: Hematoxylin and eosin stain, Klüver-Barrera, elastic Masson-Goldner, and periodic acid-Schiff. In addition, selected sections were immunostained with antibodies against β-amyloid (monoclonal; clone 6F/3D; Dako; Glostrup, Denmark; 1:100), glial fibrillary acidic protein (polyclonal; Dako; 1:1500), leukocyte common antigen (monoclonal; clone PD7/26; Dako; 1:50), CD3 (monoclonal; clone F7.2.38; Dako; 1:50), CD8 (monoclonal; clone C8/144B; Dako; 1:100), and TREX1 (polyclonal; HPA035437; Sigma Aldrich; St. Louis, MO; 1:100, after autoclaving for 10 minutes at 121°C in 10 mmol/L sodium citrate buffer). Bound antibodies were visualized by the peroxidase-polymer-based method using a Histofine Simple Stain MAX-PO kit (Nichirei, Tokyo, Japan) with diaminobenzidine as the chromogen. Immunostained sections were counterstained with hematoxylin.

To test the specificity of the antiTREX1 antibody, we performed immunohistochemistry on paraffin-embedded tissue sections of the frontal and occipital lobes, brainstem, kidney, and iliopsoas muscle of 2 individuals without any neurological disorders (2 men aged 76 and 89 years) as normal controls, and other patients with acute cerebral infarction (a woman aged 87 years) and cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy ([CADASIL] a man aged 62 years) as disease controls. We then confirmed the immunostaining of the perinuclear area and cytoplasm in the controls, as reported previously for cultured cells (6).

An electron microscopy study of the autopsied brain tissue was also performed. Formalin-fixed small tissue blocks of the grossly normal area and soft lesion in the frontal white matter were postfixed with 1% osmium tetroxide, dehydrated through a graded ethanol series, and embedded in Epon 812. Ultrathin sections were then cut and stained with uranyl acetate and lead citrate. Ultrathin sections were cut and examined with a Hitachi H-7100 electron microscope at 75 kV.

Genomic DNA and total RNA were extracted from the patient’s fresh-frozen frontal white matter. Mutational analysis of TREX1 was performed by Sanger sequencing of all coding regions, using originally designed primers: Forward primer 5′-ACACTGGGCACTCACAC-3′ and reverse primer 5′-TGACCACTCAGTGCTATGG-3′. Complementary DNA (cDNA) was synthesized with a high-capacity cDNA reverse transcription kit (Applied Biosystems; Foster City, CA). We amplified the TREX1 cDNA with originally designed primers: Forward primer 5′-TGCCTGCTACGGCTCAGC-3′ and reverse primer 5′-CCAATTGACCACTCAGTGCTATGG-3′.

Protein extracted from fresh-frozen samples of the patient’s frontal white matter was used as described previously (10). As controls, frozen tissue samples from 2 patients without any neurological disorders (2 men aged 51 and 16 years), 2 patients with Alzheimer disease (2 men aged 60 and 78 years), and 1 patient with CADASIL (a man aged 62 years) were used. Tissue lysates were separated in 2% sodium dodecyl sulfate polyacrylamide gel and subjected to immunoblotting using the antiTREX1 antibody (polyclonal; Proteintech; Rosemont, IL; 1:500) and an antiβ-actin antibody (monoclonal, Medical and Biological Laboratories Co., Aichi, Japan; 1:5000).

RESULTS

The external surface of the brain appeared normal. On gross examination, fixed sections demonstrated multiple, coalesced, and partly necrotic lesions with calcification occupying the white matter of the left and right frontoparietal lobes (Fig. 1A). Similar, but smaller lesions were also present in the right middle cerebellar peduncle. The cortical gray matter was clearly spared. Histopathology of the lesions showed confluent foci of coagulation necrosis surrounded by reactive gliosis in the white matter (Fig. 1B, C). Adjacent to the necrotic cavities, abundant macrophages, focal calcifications, and gliosis, indicating both old and fresh lesions, were evident (Fig. 1D). Injured astrocytes showing swollen and vacuolated cell bodies with beading processes (i.e. clasmatodendrocytes (11)) were also increased (Fig. 1E) in a rim of tissue surrounding the severely ischemic core. In addition, inflammatory cells consisting mainly of CD3-positive (data not shown) and CD8-positive lymphocytes had infiltrated (Fig. 1F, J). In small vessels within the lesions, fibrous expansion of the adventitia with relative preservation of smooth muscle cells (Fig. 1G) and thickening of the intimal layer (Fig. 1H) were prominent features, whereas much rarer granular calcification deposits were also evident in the walls of several vessels (Fig. 1K). Fibrinoid necrotic capillaries were seen around the large necrotic lesions (Fig. 1L). Ultrastructurally, capillaries had thickened walls with marked edema (Fig. 1I). These features were seen only around the lesions in the white matter of the frontoparietal lobes and the right middle cerebellar peduncle. Vascular thrombosis was rare. There was no evidence of vascular amyloid deposition or the extravascular granular periodic acid-Schiff-positive deposits specific to CADASIL, or vasculitis. The leptomeninges, extraparenchymal vasculature, and cortical gray matter vessels were spared.

FIGURE 1.

Histopathologic features of the frontal lobe lesion. (A) Fixed section demonstrates a large area of gelatinous necrosis in the left frontal lobe white matter and corona radiata with central cores of brownish and whitish parenchyma (arrow). Large space-occupying lesion in the left lobe white matter with central cores. (B) Calcification (arrow) close to the central core. (C) Microscopy of the central core shows gliosis and necrosis, and hyalinized vessels with massive granular calcification, corresponding to coagulation necrosis, and the appearance of an old lesion. (D) Some confluent foci of ischemia consist of infiltrated macrophages, gliosis, and calcification, suggesting that both old and fresh lesions are intermingled. (E) Abundant clasmatodendrocytes, showing swollen cytoplasmic body with beading and fragmentation of their dendritic processes (inset) are seen in the areas bordering coagulative necrosis. Several lymphocytes have infiltrated into the lesions (F), and are positive for CD3 (data not shown) and predominantly for CD8 (J). Adventitial thickening with apparently normal endothelial and smooth muscle cells (G), and thickening of the intimal layers (H) can be seen in small vessels in the ischemic regions. (K) Granular calcifications are also evident in the vascular wall. (L) Capillaries appear swollen and necrosis is evident in the fresh lesion. (I) Ultrastructure of a capillary in the ischemic region. The basement membrane is thickened with marked edema. Scale bars: A, B = 1.5 cm; C = 100 μm; D, E, F, J = 50 μm; G, H, K = 20 μm; L = 3 μm; I = 1.5 μm. Klüver-Barrera (B), H&E (C, D, F–H, K, L) and CD8 (J).

Open in new tabDownload slide

We confirmed the expression of mutant TREX1 mRNA using the patient’s frozen brain tissue, indicating that mRNA transcribed from the mutant allele was not degraded by nonsense-mediated mRNA decay. Mutational analysis of TREX1 revealed heterozygous c. 742_745dupGTCA (p. Thr249fsX14), which has been previously reported in RVCL patients with HERNS (4, 6) (Fig. 2A). Abundant wild-type TREX1 expression was evident in all tissues from the controls and the present patient, whereas expression of the truncated TREX1 protein was detected only in this patient (Fig. 2B).

FIGURE 2.

(A) mRNA expression and gene analysis of TREX1. We obtained cDNA without genomic DNA contamination from the patient’s brain by reverse transcriptase-PCR (upper column). Sequence analysis of the amplification products shows the presence of the c.742_745dupGTCA, indicating that the truncated TREX1 resulting from the mutation did not undergo nonsense-mediated decay (lower column). (B) TREX1 expression in autopsied tissue. Truncated TREX1 protein (28 kDa) is detectable in the patient, whereas no such expression is observed in the controls. (C–N) TREX1 immunohistochemistry. Images of the frontal white matter lesion (C, D, E, H), neurons of the frontal cortex (F, I), endothelial cells in the frontal white matter (G, J), the kidney (K, L), and the tubular epithelial cells (M, N). Positive reactivity for TREX1 is evident in the patient, but absent in the control (C, D). A higher-magnification view of the frontal white matter demonstrating reactivity in the oligodendrocytic nuclei in the patient (E, H). Positive nuclear reactivity in the patient (F, G) and reactivity in the perinuclear area and cytoplasm in the control (I, J). Note the difference between the patient and the control in numbers of TREX1-positive cells (arrow) in the glomerulus (K, L), and in localization of TREX1 in tubular epithelial cells (M, N). RT, reverse transcription; Con, control; Pt, patient. Scale bars: C, D = 75 μm; K, L = 50 μm; B, F, M, N = 30 μm; E, H, G, J = 20 μm.

Open in new tabDownload slide

We used an antiTREX1 antibody targeting the N-terminal domain of the protein. TREX1 immunohistochemistry demonstrated numerous positive nuclei in the lesion of the frontal white matter, most of which appeared to be oligodendrocytes on the basis of morphology (Fig. 2C, E), whereas in controls, fewer TREX1-positive cells were observed in the white matter (Fig. 2D, H). In the white matter distant from the affected areas, only a small number of TREX1-positive cells were observed. Various types of cells in the central nervous system and other systemic organs showed homogeneous immunostaining of the nuclei, or the nuclei and cytoplasm, in the patient (Fig. 2F), and of the perinuclear area and cytoplasm in the control (Fig. 2I). TREX1-positive cells were also observed among vascular endothelial cells (Fig. 2G, J). Similarly, in the kidney, abundant TREX1-positive cells with aberrant intranuclear localization, including endothelial cells, mesangial cells and podocytes, were seen in the glomeruli of the patient (Fig. 2K, M), whereas no TREX1 reactivity was seen in any glomerular cells in the controls (Fig. 2L, N).

DISCUSSION

We have described the pathologic and molecular features in the brain of an autopsied patient with RVCL harboring the T249fs mutation in TREX1. Moreover, we demonstrated for the first time the loss of perinuclear localization of endogenous TREX1 truncated proteins in the affected tissue, which had been previously reported in cultured cells (5, 6), and the distribution of the TREX1-positive cells.

The clinical presentation in the present patient resembled that of a previously reported Chinese-American case with the same mutation (4, 6) and other HERNS patients (12), all of whom exhibited visual disturbance, stroke-like hemiparesis episodes and proteinuria as young adults. Neuroradiological examination had revealed brain tumor-like lesions surrounded by extensive perifocal edema with contrast-enhancement.

The histopathology of the present case shared characteristic features of the brain and renal lesions reported in cerebroretinal vasculopathy (3), HERNS (4, 12), and hereditary systemic angiopathy (5) cases. In all cases, necrotic cavities in the white matter and confluent calcium deposits were evident. Additionally, vasculopathy, including hyalinosis and marked adventitial fibrosis with inflammatory cell infiltration, was a common feature.

It has often been reported that the brain lesions have foci of calcification that are thought to arise through a dystrophic process adjacent to areas of necrosis. On the other hand, no previous reports have described vascular calcifications in the brain of RVCL patients, although this feature has often been described in Aicardi-Goutières syndrome (AGS) (13), which is another hereditary disease caused by TREX1 mutation. AGS and RVCL share some clinicopathological features such as leukoencephalopathy with calcifications. In an in vitro model of the microangiopathy in AGS, IFN-α was reported to enhance the vascular smooth muscle cell-derived calcification (13). In RVCL cases, type I IFN activation has been reported recently (14). It is still controversial whether the type I IFN signature is a distinct feature of RVCL. However, vascular calcification is a common feature in AGS and was evident in the present patient, suggesting that these TREX1-related disorders may have overlapping pathogenetic features. Although CT in the present case demonstrated a tumor-like lesion with severe edema and contrast enhancement, the films were lost and therefore cannot be shown here. However, the presence of inflammatory cell infiltration with marked edema in the brain lesions suggests that the diseases process may include some form of immunological disorder. The possible influence of TREX1 on this immune mechanism, including regulation of interferons, warrants further investigation.

We confirmed that expression of truncated TREX1 protein and diffuse nuclear localization of TREX1 immunostaining were features exclusive to the present patient, suggesting that mutated TREX1 protein may be associated with translocation to the nucleus, as reported previously (5, 6). It has been proposed that endothelial cell dysfunction followed by disruption of the vascular basal membrane may be responsible for the systemic small vessel vasculopathy in RVCL (15). Consistent with this, we found that TREX1 was expressed and translocated into the nuclei of endothelial cells in the affected brain and podocytes, whose processes cover the glomerular basal lamina and are involved in regulation of glomerular filtration. Additionally, abundant TREX1-positive oligodendrocytes with diffuse nuclear staining were observed in the ischemic lesions of the patient. In contrast, we failed to detect any TREX1-positive oligodendrocytes or other cell types in acute ischemic lesions in the white matter of the control patient, and in the affected white matter of another patient with CADASIL. Moreover, we observed clear TREX1 positivity in several nuclei of oligodendrocytes located in the undamaged white matter of the present patient with RVCL. These data appear to be consistent with the notion that nuclear localization of TREX1 in the patient with RVCL was associated with intrinsic cellular injury rather than being a consequence of an ischemic response. Thus, it appears that mutated TREX1 in oligodendrocytes and endothelial cells may be closely related to white matter degeneration, and that podocyte and endothelial cell dysfunction may be responsible for the glomerular lesion.

To date, there has been only one report describing endogenous TREX1 expression in an autopsied brain (16). In that case, expression of TREX1 was limited to CD68-positive and Iba1-positive microglia and was upregulated in ischemic lesions, although there was no reference to mislocalization of the truncated protein. However, the cell types showing TREX1 expression did not concur with our present observations of widespread TREX1 expression throughout the brain and other organs. Because TREX1 is thought to be essential and expressed ubiquitously in mammalian cells, its expression in a wide variety of cells in both the control and RVCL would be reasonable. On the other hand, the TREX1 upregulation in ischemic lesions reported in that previous case was compatible with our present findings, and this phenomenon was also evident in the affected renal tissue.

In conclusion, the present study has revealed the clinicopathological features of an autopsied Japanese patient with RVCL, and aberrant location of the mutated TREX1 protein, suggesting the possible contribution of glial cells, including oligodendrocytes, to the white matter pathology, as well as vasculopathy. When encountering patients with tumor-like subcortical contrast-enhanced lesions, TREX1 immunostaining of biopsy samples appears to be helpful for diagnosis of RVCL. Further studies are warranted to clarify the link between mislocalization of TREX1 and the pathomechanism of systemic vasculopathy.

This study was supported in part by Japan Society for the Promotion of Science (JSPS) grants-in-aid for scientific research to Rie Saito (Award Number: 18H06211) and Taisuke Kato (Award Number: 18K07522).

The authors have no duality or conflicts of interest to declare and no funding received.

REFERENCES