Mevalonate Kinase Deficiency: A Cause of Severe Very-Early-Onset Inflammatory Bowel Disease

Abstract

Introduction

Mevalonate kinase deficiency (MKD) is a rare autosomal recessive autoinflammatory syndrome caused by mutations in the gene encoding mevalonate kinase (MVK; OMIM*251170), an enzyme involved in the biosynthesis of cholesterol and isoprenoids. It manifests as a continuous spectrum of clinical features ranging from recurrent febrile attacks associated with inflammatory manifestations, known as hyper immunoglobulin D syndrome (OMIM#260920), to a more severe form, known as mevalonic aciduria (OMIM#610377).^{1, 2} In hyper immunoglobulin D syndrome, febrile attacks are frequently associated with mild digestive features, notably abdominal pain (62% to 85%), vomiting (44% to 70%), and diarrhea (65% to 86%).² Inflammatory bowel disease (IBD)-like involvement has been observed in 16% of the 114 patients with MKD included in the EUROFEVER registry² and some case reports.^{3, 4} However, data on MVK-associated IBD remains limited. Herein, we aim to describe the spectrum of the clinical, pathological and course of MKD-related severe IBD-like features in a series of 10 patients to improve diagnosis and management of this rare but severe condition.

PATIENTS AND METHODS

Patients

The inclusion criteria for this French retrospective multicenter study required the presence of biallelic mutations of MVK and severe IBD-like features defined by the need for abdominal surgery and/or parenteral nutrition at diagnosis or during the course of MKD. All patients in whom biallelic MVK mutations had been indicated and who were followed in French pediatric centers between January 1999 and June 2014 were eligible. Referring physicians were invited by mail to participate in this survey by filling out a questionnaire sent online. We recorded demographic data, clinical manifestations, laboratory evaluations, gastrointestinal endoscopic and histologic findings, and treatment efficacy and tolerance. The same pathologist (DB) centrally reviewed all available digestive biopsies. Complete remission of MKD-related IBD-like features was defined as the disappearance of all clinical digestive manifestations, parenteral withdrawal, weight improvement, and normalization of serum concentrations of albumin, C-reactive protein (CRP), erythrocyte sedimentation rate, and fecal calprotectin during follow-up.

Genetic Studies

Sanger sequencing of MVK was performed as previously reported.⁵ In addition, when DNA was available, targeted next-generation sequencing was used to analyze 67 genes in which mutations can cause chronic diarrhea and very-early-onset IBD (VEO-IBD).⁶ Informed consent for genetic analysis was obtained from all patients or their parents. The Immunobiota study (study for investigation of VEO-IBD by NGS) was approved by the Comité de Protection des Personnes (identification number for registration in France [ID-RCB]: 2014-A00017-40, Agence nationale de sécurité du médicament et des produits de santé [ANSM] reference: 140044B-42).

RESULTS

Ten patients from 8 centers were included. Two of these patients were previously reported (patients 2 and 5).³ Demographic features and clinical endoscopic and histologic features of severe IBD-like manifestations are shown in Table 1 and Supplemental Table 1. Age at onset of MKD ranged from 1 day to 17 months. Febrile attacks were present in all patients and preceded IBD-like features in 4 patients (40%). Genetic analysis revealed homozygous (1 patient) or compound heterozygous (9 patients) genomic mutations in MVK (Table 2). The median duration of follow-up from the diagnosis of IBD was 9 years (range, 4-20 years).

Age at onset of IBD manifestations ranged from 1 day to 6 years (median, 3 months). These manifestations revealed MKD in 5 patients (50%), developed within the first 2 years of life in 7/10 patients (infant-onset IBD), and before age 6 years in all patients (VEO-IBD). All IBD-like features presented as abdominal pain, associated with diarrhea (7 patients) and/or vomiting (9 patients), and septic shock (1 patient). Physical examination revealed acute abdomen (6 patients) and anal fistula (2 patients). All patients had elevated acute phase reactants and IBD inflammation parameters: median CRP 86 mg/L, median fecal calprotectin 889 µg/g, median albumin 30 g/L, and median hemoglobin 9 mg/dL. Seven patients underwent a colonoscopy, which showed colitis in 6 patients (86%) that was associated with mild to profound ulcerations intermingled with patches of healthy mucosa (3 patients). A clinically suspected rectal fistula was confirmed after examination under general anesthesia or colonoscopy in 2 patients (patients 7 and 5, respectively). Surgery was performed in 7 patients because of either acute abdomen (6 patients) or drainage of rectal abscess (1 patient). Laparotomy showed abdominal adhesions (6 patients), bowel perforation (1 patient), ileal necrosis (1 patient), and sterile peritonitis (1 patient; Table 1).

Histologic analysis of ileal and colonic biopsy samples obtained from the sites of inflammation (n = 6) showed subtotal to total small intestine villous atrophy (jejunum or ileum) without increase of intraepithelial lymphocytes and moderate to profuse lamina propria mononuclear inflammatory infiltrates containing lymphocytes, plasma cells, and, more particularly, numerous neutrophils with cryptic abscesses. Other frequent histologic features were profound and large ulcerations, that were inconstantly associated with necrosis and/or perforation and abundant glandular apoptosis.

Complete remission was achieved in all patients, with a median duration of follow-up of 3.3 years (range, 0.5-8 years). At last time of follow-up, all inflammation parameters were normalized: CRP (86 mg/L vs 0 mg/L), albumin (30 g/L vs 41 g/L) and hemoglobin (9 mg/dL vs 11.8 mg/dL). Fecal calprotectin was not available at the last time of follow-up. Because clinical remission was achieved for a long period, fecal calprotectin was not available at the last time of follow-up. There was no difference in median body mass index (14.6 vs 14.7) between diagnosis and clinical encounter. At the time of IBD onset, no patient had received biologics. In all 7 patients who underwent surgery, medical treatment was required afterward to achieve remission: corticosteroids alone (3 patients) or associated with an anti-interleukin (IL)-1 agent (anakinra or canakinumab, 2 patients), or an anti-IL-1 agent alone (2 patients). In the remaining 3 patients who presented with inflammatory colitis and did not require abdominal surgery, anti-IL-1 agents allowed sustained remission with rapid tapering of corticosteroids in 6 months or less. After initiation of anti-IL-1 therapy, clinical remission was respectively achieved within a median of 5 days and 1 month (median dosage of anakinra 4 mg/kg/day and canakinumab 2.3 mg/kg/month), and anti-IL-1 therapies were still ongoing at last follow-up in 7 patients (Table 1). The 3 remaining patients were lost to follow-up, and 2 still required steroids. Conversely, anti-TNF agents were inefficacious in 2 patients, with patient 5 rapidly (1 month after initiation) developing auto-antibodies against infliximab and anti-TNF-induced systemic lupus erythematosus. Ileostomy, colostomy, and/or surgical intestinal resections were needed in 2 patients (Table 1). Bowel continuity was restored 1.5 and 9 months after surgery in patient 2 and patient 4, respectively, while they received corticosteroids and anti-IL-1 agents. Five patients required total parenteral and/or enteral nutrition for 0.5 to 97 months. Three months after beginning anakinra dosage, a colonoscopy showed complete mucosal healing and the disappearance of stenosis and fistula in patient 5; patient 10 had also a normal colonoscopy after anakinra initiation.

Targeted next-generation sequencing was performed in 8 patients and confirmed MVK mutations but did not identify other VEO-IBD-causative monogenic mutations. Notably, variants in genes associated with autosomal recessive diseases were observed in 7/8 patients but only at the heterozygous state. With the exception of SLC2A2 (n = 1) and ApoB (n = 2), which were predicted to be pathogenic, all variants were expected to be benign polymorphisms (Table 2).

Discussion

The recognition of VEO-IBD as a feature of MKD has important implications for clinical management. Indeed, all of the patients in our study met the definition of VEO-IBD, and MKD was diagnosed after observing initial IBD-like features in half of our cohort. Moreover, our study provides a detailed description of IBD-like features associated with MKD and reports for the first time that histologic analysis may contribute to the diagnosis of MKD. Similar to our observations, the Eurofever cohort and some case reports have also described MKD-related IBD, such as aseptic peritonitis, gastrointestinal bleeding, intestinal occlusion, gut perforation, abdominal adhesions, and gastrointestinal or perianal ulcers.^3-5 The majority of patients presented with perianal fistula and abdominal adhesions, which likely resulted from repeated acute sterile peritonitis.²

Notably, our report also showed a severe intestinal inflammation mimicking VEO-IBD in 4 patients before recurrent fever onset. In addition to conventional IBD, comprising complex multifactorial disorders, IBD-like intestinal inflammation has been linked to an expanding number of rare monogenic disorders.^{6, 7} Although penetrance varies, intestinal inflammation generally develops before age 6 years (all patients in our cohort), defining VEO-IBD.⁷ A causative role of MKD in severe intestinal inflammation was first suggested by an earlier report of 2 patients in whom severe neonatal-onset colitis revealed the genetic defect.⁴ None of our 7 tested patients carried other homozygous mutations known to be responsible for monogenic IBD. These data suggest that homozygous or compound heterozygous MVK mutations are a genetic causative defect of VEO-IBD. However, considering that variants predicted to be deleterious were found at the heterozygous state in genes responsible for autosomal recessive IBDs in some patients, we may not rule out that MVK mutations may also interfere with the expression of other mutations predisposing to IBD in some patients, as previously suggested.⁸

As the majority of monogenic IBD,⁷ MKD-related IBD presented as VEO-IBD, with a high proportion of severe colitis and ulcerating and fistulizing disease. Centralized histologic analysis showed that gut inflammation associated with MKD did not classify as Crohn’s disease or ulcerative colitis, according to accepted histologic criteria. Moreover, MKD-related IBD pathological description is also different from that of autoimmune enteropathies. Therefore,we suggest considering MKD diagnosis when VEO-IBD-like disease is associated with adhesions, rectal or anal fisulae and enterocolitis, and the following histologic features are found: small intestine villous atrophy without intraepithelial layer increase, lymphoplasmocytic and especially neutrophil inflammatory infiltrate with cryptic abscesses, profound and large ulcerations inconstantly associated with perforation and/or necrosis and/or glandular apoptosis.

Anti-tumor necrosis factor-alpha treatment did not allow remission in the patients in our cohort. In contrast, anti-IL-1 therapy was efficacious in all 7 patients with MKD-related IBD. An increased production of IL-1β in patients with MKD represents a putative link between MKD and inflammation,⁹ which is supported by the efficacy of anti-IL-1β agents in MKD-related manifestations.¹⁰ Thus, anti-IL-1β blockade should be considered in patients with VEO-IBD who are unresponsive to conventional drugs.

Conclusions

Our study suggests that MKD should be considered as a cause of monogenic VEO-IBD, especially in patients with peritoneal adhesions, perianal disease, and/or atypical IBD pathology and/or a history of recurrent or chronic fever. Anti-IL1β therapy may be efficacious in these patients with MKD-related VEO-IBDs.

Author Contributions

BBM and IM designed the study. BBM, ALM, AB, FU, VRP, SGL, and IM collected clinical data. LC, FCH, and NCB collected genetic data. BBM, DB, LC, FCH, NCB, and IM analyzed data. BBM, ALM, and IM wrote the article. BBM and IM supervised the study. All authors have read final approval of the version published.

Supported byIM was supported by a grant from the Bettencourt Schueler Foundation through the Imagine MD-PhD program.

Conflicts of InterestBBM has received financial support to attend congresses from Novartis and Sobi and has participated as a coinvestigator in clinical trials by Novartis and Pfizer but not for this work. PQ has received consultancy or speaking fees from AbbVie, Bristol Myers Squibb, Chugai-Roche, Lilly, Novartis, Novimmune, and Swedish Orphan Biovitrum, but not for this work; he has also participated in a data safety monitoring board for Sanofi. FR received research support or consultancy fees from AbbVie, Nestlé, Pfizer, Celgene, MSD, Johnson & Johnson, Janssen, and Takeda. FU has received financial support to attend congresses from Novartis. SGL has received counseling or travel expense fees from Novartis and SOBI in the past, but not for this work. IM has received counseling fees or travel expense fees from Novartis and SOBI in the past, but not for this work. ALM, LC, FCH, NCB, JV, AB, VRP, UM, AF, and DB report no conflicts of interest.

Data Availability Statement

The data underlying this article will be shared upon reasonable request to the corresponding author.

References

Author notes