Microvillus inclusion disease,a diagnosis to consider when abnormal stools and neurological impairments run together due to a rare syntaxin 3 gene mutation

1. Introduction

Microvillus Inclusion Disease (MVID) was first described in the literature in 1978 with presentation of severe watery diarrhea, failure to thrive, and metabolic acidosis [1]. Intractable watery diarrhea typically presents within the first days of life, however there is a milder form of MVID with onset after two to three months of life. Stool volumes can range between 150 to 300 ml/kg/day and causes severe dehydration and bicarbonate loss leading to marked metabolic acidosis [1, 2]. Infants continue to have large-volume stools despite bowel rest. Typically, there are no additional clinical signs or organ system involvement outside of the gastrointestinal tract apart from cholestatic liver disease in approximately one-third of patients [1, 2].

Diagnosis is made by analysis of small intestinal biopsies by electron microscopy showing characteristic findings of microvillus inclusions in the cytoplasm of enterocytes with shortening or absence of microvilli on the apical or basolateral surface, and the apical accumulation of electron-dense granules. Staining with periodic acid-Schiff (PAS) may also show subapical accumulation of secretory granules [2]. Treatment consists of fluid and nutrient supplementation typically requiring life-long TPN and small bowel transplantation [1–3]. Although prognosis remains poor despite medical treatment, early identification and prompt referral to a transplant center can optimize nutrition and lifesaving intestinal transplant can be done in a timely manner [3]. Mutations in the myosin Vb (MYO5B) gene have been identified as causative for MVID [4, 5]. Recently, a second gene was identified (STX3) with only four reported cases in the literature and phenotype limited to intestinal disease [1, 6–8]. We report a fifth patient with a novel STX3 mutation with MVID-like phenotype and neurological involvement.

2. Methods

Data regarding the patient’s clinical course, laboratory results, and imaging results was obtained through retrospective chart review of the patient’s electronic medical record (EMR). Documents including the admission history and physical, daily progress notes by the patient’s primary team and consulting services, and the discharge summary were individually reviewed.

Biopsies of the stomach, esophagus, and duodenum, as well as the sigmoid colon and rectum, were obtained during esophagogastroduodenoscopy (EGD) and flexible sigmoidoscopy, respectively. The samples were sent to the department of pathology at University of Texas Health San Antonio for review by light and electron microscopy.

3. Case presentation

A term male of Afghan decent was born to consanguineous parents via repeat C-section with Apgar scores of 7/9 at 1, 5 minutes respectively. Of note, the parents had a previous infant die on day of life (DOL) 4 following an abdominal surgery for what is presumed to be imperforate anus. The current pregnancy had been complicated by polyhydramnios, pre-eclampsia, and chorioamnionitis. Exam was remarkable for abdominal distension concerning for obstruction, for which the infant was kept non per os (NPO) and admitted to the neonatal intensive care unit (NICU). Abdominal x-ray showed scant bowel gas but was otherwise unremarkable, abdominal ultrasound showed meconium filled bowel loops, and a lower GI contrast study showed normal appearing rectum, sigmoid, and proximal descending colon with meconium plugs present. The patient passed meconium on DOL 1 and feeds were advanced. Due to concerns for GI malformations, echocardiography and echoencephalography were performed to rule out other anomalies, and were normal.

On routine lab monitoring on DOL 6 the infant was found to have severe metabolic acidosis (HCO3 12 mEq/L and anion gap of 14 mEq/L). He underwent a septic workup and abdominal x-rays revealed dilated bowel loops without air in the rectum. He was made NPO and TPN was initiated. An upper GI with gastrografin ruled out malrotation and obstruction. Gastroenterology was consulted given high clinical suspicion for Hirschsprung’s disease, which was ruled out by suction rectal biopsy. Acidosis corrected with TPN and feeds were resumed on DOL 10, and slowly advanced. On DOL 17, with a feeding volume of approximately 100 mL/kg/day, the patient became severely acidotic (HCO3 11 mEq/L and anion gap of 11 mEq/L), feeds were held and TPN initiated. This cycle recurred despite multiple changes in enteral nutrition to include hydrolyzed and amino acid based formulas, and ultimately the infant only tolerated pedialyte.

Throughout his hospitalization his clinical course was remarkable for poor oral feeding, non-bilious non-bloody emesis, low-volume diarrhea with stools of varying consistency (mucoid string-like stool or crystals scattered throughout yellow gelatinous stool). Initial newborn neurological exam was normal but by two months of age he developed an abnormal neurological status (poor response to sound, intermittent nystagmus, staring spells, and tone fluctuation between hypo-and hypertonia). Neurological exam showed positive light perception but lack of tracking and at times no downward movement of the eyes. There was normal bulk but increased muscle tone in bilateral upper greater than lower extremities, and decreased truncal tone. Deep tendon reflexes were found to be normal. Neurologic workup with MRI of the brain, ophthalmologic examination, and audiologic evaluation with auditory brainstem response testing were normal.

GI imaging series, metabolic work-up, and cystic fibrosis work-up were all normal. Stool studies showed normal pH, trace reducing substances, normal pancreatic elastase and fecal fat consistent with malabsorption. Fecal osmolality was elevated. At six weeks of life he underwent his first esophagogastroduodenoscopy/sigmoidoscopy which showed grossly normal duodenal mucosa and non-specific colitis in the sigmoid colon (Figs. 1a and 2a) Biopsies obtained revealed features suggestive of MVID on electron microscopy evaluation of a duodenal biopsy with only one intracytoplasmic pocket of microvilli. Repeat biopsies taken at 10 weeks of life showed irregular surface microvilli with some areas lacking microvilli and apical electron dense secretory granules present in the duodenum (Fig. 1c), and loss of surface microvilli, apical secretory vacuoles, and atypical inclusions along the basolateral cellular borders in the sigmoid colon (Fig. 2b-c) On light microscopy of the duodenum, H&E morphology was unremarkable (Fig. 1b) Immunohistochemistry (IHC) of the sigmoid colon with CD10 showed preserved brush border with staining of apical cytoplasm and rare cytoplasmic globules (Fig. 2b) On genetic testing, no MYO5B gene variants were identified, however whole exome sequencing was positive for a homozygous loss of function gene mutation at STX3 (c.363_366delinsGA; p.Val122fs), confirming the diagnosis of MVID.

OPEN IN VIEWER

Fig. 1.

a. Endoscopy image of normal appearing duodenum. It is important to note that the pathogenesis of MVID is due to a functional defect on the electron microscopy level, so normal looking mucosa does not rule out the diagnosis. b. Duodenal biopsy showing normal histology with preserved villous architecture (H&E, original magnification×100). Findings in MVID can be patchy, so taking multiple biopsies is important for not missing the diagnosis. c. Duodenum showing subapical collections of dense granules and focal loss of surface microvilli (arrow) (Electron Microscopy, original magnification×800).

OPEN IN VIEWER

Fig. 2.

a. Endoscopy of Sigmoid Colon showing non-specific colitis, which is a distraction from the real diagnosis in this case. b. CD 10 IHC showing preserved brush border with increased apical cytoplasmic staining and rare intracytoplasmic globules in the sigmoid colon (original magnification×400). c. Sigmoid colon showing two atypical inclusions located along the lateral walls of the enterocytes in the basal portion of the enterocyte and above the basement membrane. Cross sections of the villi contain filamentous cores (red arrows show inclusions; yellow arrow denoting basement membrane) (original magnification×6,000).

Microvillus inclusion disease is a rare disease with autosomal recessive inheritance, and is more common in populations with a high rate of consanguinity [1, 4]. The estimated prevalence is unknown, with sources reporting at least 200 cases in Europe, however the disease occurs worldwide [4]. Mutations in the MYO5B gene were identified in 2008 as causative for MVID [4], and one study reports that gene sequencing and multiplex ligation-dependent probe amplification analysis detect biallelic mutations in >90% of patients with a clinical and histopathologic diagnosis of MVID [1].

The second gene identified as causative for variant MVID was syntaxin 3, reported in 2014 by Wiegerinck CL, et al., in two cases in which MYO5B mutations were negative by Sanger sequencing [6]. Whole exome sequencing was performed independently for these patients, resulting in theoretically disease-causing mutations. The first patient was a Dutch newborn who presented with watery diarrhea and severe metabolic acidosis on the second day of life, her mutation was identified as a homozygous nonsense mutation in exon 9 (c.739C > T; p.Arg247Ter). The second patient was a boy from Pakistan who presented with frequent watery stools in the second week of life, several episodes of severe acidosis, frequent pulmonary infections, renal tubulopathy, and episodic vomiting. In this patient, a homozygous mutation causing frame-shifting insertion in exon 6 of STX3 was identified (c.372_373dup; p.Arg125fs). Both mutations resulted in STX3 protein depletion and truncation [6]. Neither of these patients have been reported to have neurologic abnormalities.

Two other patients with MVID phenotypes and mutations in STX3 are also reported in the literature. Alsaleem et al. reported a male Saudi infant born to a non-consanguineous couple who presented with severe vomiting, mild diarrhea, weight loss, and metabolic acidosis starting on DOL 3. He was found to have a homozygous nonsense gene mutation in STX3 (c.739C > T; p.Arg247*) [8], the same mutation reported by Wiegerinck et al. [6]. Despite the same genotype, this patient was reported to have severe vomiting but only mild diarrhea, and was able to tolerate some enteral feeds with elemental formula [8]. Chen et al. report a female infant presenting with yellow-green mucusy diarrhea and metabolic acidosis on the second day of life. Her sequencing results showed two nonsense mutations of STX3 (c.424C > T; p.Arg142 and c.739C > T; p.Arg247), which were inherited from her parents who showed normal phenotype [9].

Interestingly, there has been one non-MVID patient reported with a mutation in STX3. This patient who was born to a consanguineous Tunisian family, was found to have a novel missense mutation c.122A > G; p.E41G resulting in the phenotypic findings of congenital cataracts, intellectual disability, and psychomotor retardation without intestinal symptoms [10]. STX3 is highly expressed in the brain and retina [10], which explains why both organs are usually affected, however the absence of intestinal symptoms in the patient of Tunisian descent is not well explained, and further investigations are necessary.

Our patient, who is the first one to be found to have a combination of MVID and neurologic abnormalities, had a novel frameshift mutation at codon 122, truncating the 289-amino acid protein and likely resulting in loss of function (c.363_366delinsGA; p.Val122fs). The c.363_366delinsGA variant has not been previously identified in general population frequency databases (1000 Genomes, Exome Sequencing Project, Exome Aggregation Consortium (ExAC)). Truncating variants in this gene are very rare and have not been documented in the homozygous state in ExAC. Therefore, this particular mutation might express both, intestinal and neurological phenotypes more often and the clinician must be aware of symptomatology.

It remains to be determined why this infant without copious diarrhea would develop such severe metabolic acidosis. We speculate these abnormalities were due to a severe malabsorptive capacity of bicarbonate in particular without increased water losses. The pathophysiology of MVID is only partially understood. Loss of function of the MYO5B protein results in disruptions in protein trafficking in enterocytes, leading to mistargeting of apical and basolateral proteins. This disrupted trafficking is thought to cause a defect of epithelial cell polarity, leading to loss of brush border integrity, absorption deficits and watery diarrhea seen in these patients [4]. In 2015, Carton-Garcia et al. described a novel mouse model with targeted inactivation of MYO5B. These knockout mice showed the typical clinical features of MVID as well as ultrastructural changes seen on electron microscopy [5, 11]. No knockout mouse model exists for STX3 mutations, however Wiegerinck et al. have shown that truncated versions of STX3 in Caco-2 cell cultures have recapitulated all histologic hallmarks of MVID [6]. An improved understanding of the pathophysiology of MVID will allow development of targeted therapies which could potentially lead to improved quality of life and decreased morbidity and mortality of patients with MVID.

In addition to an atypical clinical presentation for MVID, results of electron microscopy of small and large intestinal biopsies showed some, but not all, of the classical features of MVID. These features include absence, atypical, or partial loss of surface microvilli, cytoplasmic microvillus inclusions, and subapical secretory granules-typically seen in duodenal biopsies. On initial biopsy at six weeks of life, only one intracytoplasmic pocket of microvilli was noted on EM evaluation of the duodenum, which suggested MVID; however, all other classical features were absent. He underwent a subsequent biopsy at 10 weeks of life which showed slightly irregular surface microvilli with some areas lacking microvilli, and apical electron dense secretory granules in the duodenum. There were no duodenal intracytoplasmic microvillus inclusions. However, evaluation of the sigmoid colon did reveal irregular surface microvilli, apical secretory vacuoles, and atypical basolateral collections of microvilli. The appearance of microvilli located at the basolateral surface are characteristic of STX3-driven MVID patients and are not seen in MVID patients with MYO5B mutations [12]. Additionally, such inclusions have not been routinely described in the colon as most biopsies have been of the small intestine. However, Schofield et al. described one patient with MVID where microvillus inclusions were present in the duodenum, jejunum, ileum, and colon [13].

This case of an infant with low volume mucus-like stools, metabolic acidosis, and abnormal neurologic findings demonstrates a unique presentation of MVID in a patient with a rare genotype. Unlike the typical presentation, our patient never demonstrated severe, watery diarrhea yet was unable to tolerate enteral nutrition due to metabolic acidosis. Furthermore, neurological findings have never been described in MVID with this novel mutation. We speculate the neurological findings are due to abnormal trafficking in the neurons of surface receptors/proteins.

Today, the patient is 15 months old, is fully TPN dependent with small-volume pleasure feeds of Pedialyte, experiences cortical visual impairment with optic atrophy and nystagmus, hearing loss, and is developmentally delayed receiving physical, occupational, speech and visual therapies. He is followed by multiple subspecialties including gastroenterology (GI), ophthalmology, otolaryngology, hematology for refractory anemia, audiology, and pediatric surgery. He experiences recurrent inpatient admissions to the Pediatric Gastrointestinal service, mostly for complications related to his central line and TPN.

5. Conclusions

It is important to consider MVID in the differential diagnosis of a neonate/infant with abnormal stools, metabolic acidosis, with and without neurologic symptoms. Practitioners should consider early referral to GI specialist for esophagogastroduodenoscopy/sigmoidoscopy with intestinal biopsies and targeted genetic testing for this condition.

Disclosures

The authors have no financial or material disclosures.