Susceptibility of nucleotide-binding oligomerization domain 2 mutations to Whipple’s disease

Abstract

Introduction

Nucleotide-binding oligomerization domain containing protein 2 (NOD2) is an intracellular innate immune sensor that plays an important role in defence against intestinal infection and inflammatory regulation [1]. NOD2 gene mutations are associated with systemic autoinflammatory diseases, such as Crohn’s disease, Blau syndrome and Yao syndrome (YAOS, OMIM #617321), formerly designated as NOD2-associated autoinflammatory disease. The latter disease is characterized by recurrent episodes of fever, dermatitis, arthralgias, distal leg swelling, gastrointestinal complaints, sicca-like symptoms and eyelid swelling [2, 3]. Diagnosis requires a characteristic clinical phenotype and the presence of certain NOD2 mutations with exclusion of other diseases. Most patients are compound heterozygous for the NOD2 IVS8 + 158 [minor allele frequency (MAF) 5%] and another low frequency or rare NOD2 variant such as R702W (MAF 1.4%), 1007fs (MAF 0.6%), V955I (MAF 3.3%), or other rare NOD2 variants [4, 5]. This disease has recently been added to the new category of Genetically Transitional Disease [6]. It is a new concept in genomic medicine and denotes a disease or disease status straddling the monogenic and polygenic disease, where mutation is necessary but not sufficient to cause disease.

Whipple’s disease (WD) is caused by the Gram-positive actinomycete, Tropheryma whipplei (TW), and was first described in 1907 by George H. Whipple, an American pathologist. TW is thought to be transmitted through oral–oral and faecal–oral routes. In the general population, ∼75% of people clear TW infection, and asymptomatic carriers account for 2–37% [7]. The rod-shaped microorganism has a trilaminar plasma membrane that is surrounded by a homogeneous cell wall with muramyl dipeptide [8]. WD is clinically characterized by constitutional symptoms, and gastrointestinal and/or extraintestinal manifestations. In the literature, most reports have been single cases owing to the relatively uncommon nature of WD. Biopsy of tissues, often small bowel, is the diagnostic test of choice. The characteristic histopathologic findings include foamy macrophages with periodic acid–Schiff-positive (PAS+) materials. PCRs of blood, other sterile body fluids and tissues to detect bacterial 16S rRNA–specific gene can also facilitate diagnosis especially if tissue pathology is non-diagnostic. Genetic susceptibility to WD has been speculated based on familial cases, but there are few studies available [7, 9, 10].

We have not seen studies to address a possible association of NOD2 mutations with TW infection or WD. Herein, we report three patients with NOD2 mutations associated with autoinflammatory disease complicated with WD to support the contribution of NOD2 mutations to the development of WD as a host susceptible gene for TW infection.

Patients and methods

A multicentre, retrospective study of three patients with NOD2 mutations and WD was conducted. All three patients received multidisciplinary care with extensive workups at multiple institutions with records and case details collected from the time of the first evaluation among these institutions until the most recent follow-up before 1 March 2023. There were no systemic autoimmune diseases, IBD, infections other than TW or malignancies found. All authors participated in care for these patients. The medical records are available for case 1 and 2 at Stony Brook University Hospital, for case 2 and 3 at Mayo Clinic, and for case 1 at New York University Hospital. Cases are summarized below with updates on case two and three as they were published as ‘Clinical Problem-solving’ case in the New England Journal of Medicine [11] and ‘Electronic Clinical Challenges and Images in GI’ in Gastroenterology [12]. In these articles, no details on genetic testing results of patients were provided, and there were no discussions on a potential relationship between NOD2 and TW or WD. This study was approved by the Institutional Research Board of Stony Brook University. Informed consent was provided for the publication of this paper.

Results

Case descriptions

Case 1

A 35-year-old Caucasian man presented with 10 years of episodic high fevers associated with inflammatory polyarthritis affecting the hands, elbows, hips and feet, small patchy facial erythema, and eyelid swelling and discolouration. He developed intermittent loose stools and nausea and vomiting, initially without abdominal pain. Bidirectional endoscopies [esophagogastroduodenoscopy (EGD) and colonoscopy] were unrevealing in 2019. MRI of the abdomen reported mild splenomegaly and mesenteric lymphadenopathy with biopsies showing non-necrotizing granulomas. Inflammatory markers were chronically elevated. Workup, including a bone marrow biopsy, was negative for malignancy or infection. Genetic testing for FMF was negative. He was treated with empiric colchicine for an unspecified autoinflammatory disease, followed by MTX, anakinra, adalimumab and tocilizumab trials (Table 1). In June 2021, due to ongoing diarrhoea and abdominal pain, bidirectional endoscopy with duodenal biopsies was completed and demonstrated macrophage infiltrate with PAS+ staining materials confirming TW (Fig. 1). Immunosuppressive therapy was stopped. He was treated with ceftriaxone for 2 weeks followed by trimethoprim–sulfamethoxazole, which was switched to doxycycline and HCQ due to development of suspected sulfonamide drug-induced neutropenia. Due to clinical concern of autoinflammatory disease, further genetic testing using a periodic fever syndrome gene panel that includes MEFV, TNFRSF1A, NLRP3, MVK, NLRP12 and NOD2 (DDC, Middlefield, OH, USA) identified a heterozygous NOD2 mutation, V955I. Fever and gastrointestinal symptoms significantly improved with antibiotics, but intermittent inflammatory arthritis (wrist and ankle), facial rash, eyelid swelling and loose stools continued.

Figure 1.

Representative duodenal histopathology in Whipple’s disease. Duodenal villi distended by an infiltrate of foamy macrophages and admixed fat droplets, hematoxylin–eosin, original magnification ×100 (A); periodic–acid Schiff (PAS) stain showing macrophages with intense positivity, ×200 (B); high-power image of PAS stain demonstrating macrophages packed with strongly PAS-positive Tropheryma whipplei ×400 (C)

Open in new tabDownload slide

Case 2

A 68-year-old Caucasian male presented with 5 years of intermittent fever and 4 years of polyarthralgia, predominantly affecting the hips, knees and ankles. Fever often occurred at night with associated sweats. He reported more recent onset of lower extremity pain and swelling in the preceding 8 months. He endorsed mid-abdominal pain without diarrhoea along with a 22 kg weight loss and recurring pleuritic chest pain. He experienced rash with erythematous papules on his forearms. Bidirectional endoscopy (June 2021) was normal. Despite elevated inflammatory markers, there was no identifiable systemic autoimmune disease. PET-CT and bone marrow biopsy were negative for malignancy. Imaging showed bilateral small pleural effusions, pericardial effusion and benign-appearing mesenteric lymphadenopathy. He received high-dose prednisone with improvement and was maintained on prednisone 7–10 mg daily. MTX was ineffective. Genetic testing with Autoimmune and Autoinflammatory Syndrome panel (156 genes, Invitae, San Francisco, CA, USA) identified a high-risk heterozygous allele in the NOD2 gene, 1007fs. Whole NOD2 gene sequencing (DDC, Middlefield, OH, USA) reported compound NOD2 heterozygote, IVS8 + 158 and 1007fs. In March 2022, he presented with worsening dyspnoea and pleurisy. Cardiac MRI demonstrated myopericarditis with bilateral pleural effusions. Colchicine was started for pericarditis. Thoracentesis was performed. Pleural and duodenal PCR and duodenal biopsy confirmed the diagnosis of WD. The patient received one subcutaneous injection of canakinumab 150 mg plus ceftriaxone i.v. for 2 weeks followed by trimethoprim–sulfamethoxazole. His symptoms improved significantly until early July 2022 when trimethoprim–sulfamethoxazole was discontinued due to hyperkalaemia. A few weeks later, he developed right upper quadrant abdominal pain and bloating without diarrhoea. He was given ceftriaxone for 3 weeks for suspected WD relapse followed by doxycycline and HCQ. Sulfasalazine was added for continued abdominal pain, rash and low-grade fever suspected secondary to autoinflammatory disease. He then developed left periorbital swelling with erythema and was treated with 5 days of additional antibiotic for possible cellulitis with resolution. The patient’s overall symptoms significantly improved with long-term antibiotics for WD. At our request, benign genetic variant report in the Invitae lab listed a heterozygous MEFV mutation, K695R.

Case 3

A 42-year-old Caucasian man presented in 2018 for 10 years of progressively severe episodes of malaise, low-grade fever, polyarthralgia (shoulders, knees, ankles, hands and wrists) and elevated inflammatory markers. Systemic autoimmune diseases including classic CTDs and vasculitis were excluded after an extensive workup. Trial of MTX for inflammatory arthritis resulted in worsening leucocytosis, transaminitis and abdominal discomfort. Abdominal CT and PET-CT showed mildly hypermetabolic cervical and axillary lymph nodes and few prominent mesenteric and retroperitoneal lymph nodes. Infectious and haematologic evaluation were negative though endoscopies were not initially done. Mesenteric lymph node core needle biopsy was negative for malignancy. Multiple DMARDs including numerous biologics (Table 1) were tried (2017–21) without improvement in recurrent fever and night sweats. In 2020, there was one episode of erythematous coin-like rash lesions on his upper arms/chest and eyelid swelling. Low-dose prednisone achieved symptomatic improvement in malaise and arthralgia. Molecular testing for 72 genes (Invitae, San Francisco, CA, USA) identified a heterozygous NOD2 variant, 1007fs. Due to worsening gastrointestinal symptoms and weight loss in December 2021, EGD was done, and duodenal biopsy confirmed the diagnosis of WD. He completed 3 weeks of ceftriaxone followed by 1 year of trimethoprim–sulfamethoxazole. Prednisone was tapered and able to be stopped. Clinical symptoms resolved except for polyarthralgia which was treated with HCQ.

Discussion

Herein we report three Caucasian men with carriage of NOD2 mutations, as well as manifestations associated with autoinflammatory diseases, eventually complicated with WD. Our study has provided preliminary evidence of NOD2 deficiency as a host genetic risk for development of WD.

WD pathophysiology and its association with NOD2

WD is pathologically characterized by macrophage infiltrates of intestinal mucosa. The TW bacteria replicate within mucosal macrophages and peripheral blood mononuclear cells [13]. This process renders monocytes and macrophages dysfunctional, resulting in impaired degradation of intracellular bacteria with eventual accumulation of PAS+ materials or inclusion bodies [14]. Therefore, WD is considered as a macrophage-driven disease [8].

NOD2 protein predominantly resides in monocytes/macrophages. It consists of leucine-rich repeats (LRRs) at C-terminal, central region, nucleotide binding domain (NBD), and Caspase recruitment domains (CARDs) at N-terminal. NOD2 recognizes and binds bacterial peptidoglycans or muramyl dipeptide via the LRR region and activates the NOD2 signalling pathway, leading to cytokine production and subsequent inflammatory response (Fig. 2) [1]. Mutated NOD2 encoding the LRR region impairs the pathogen recognition phase and subsequent NOD2 signalling pathway activation [15]. Mice deficient for the NOD2 gene are susceptible to bacterial infection that occurs only by means of intragastric dosing rather than parenteral administration suggestive of a protective role of NOD2 against intestinal infection [16]. In addition, chronic NOD2 stimulation increases bacterial killing of monocyte-derived macrophages through upregulated reactive oxygen/nitrogen species pathway. NOD2 genetic defect in patients reduces bacterial killing [17]. NOD2 1007fs is a frame-shift mutation and encodes a truncated protein resulting in loss of function. Two of our patients carried the mutation, supporting that this confers higher susceptibility to the strain of TW bacteria. NOD2 V955I is located in exon 9–encoding protein within the LRR region, and this mutation has been associated with YAOS [5]. The LRR-associated NOD2 mutations have been known to impair bacterial recognition, though the functional impact of the NOD2 V955I has not been studied or reported. Taken together, these data highly support the role of NOD2 LRR-associated variants in TW infection and development of WD. Herein, it is important to mention the mechanism of NOD2-mediated regulation of innate immune response to intestinal microflora. The NOD2 signalling pathway involves receptor interacting protein kinase 2 (RIP2) that can be activated dependently on or independently of NOD2 [18] There is a cross-talk between NOD2 and Toll-like receptors (TLRs). NOD2 activation causes IFN regulatory factor 4 (IRF4) expression, which in turn binds to TNF receptor associated factor 6 (TRAF6) and RIP2, leading to nuclear factor-κB activation [19]. NOD2 downregulates inflammation via IRF4-mediated inhibition [19] and interestingly, an IRF4 gene mutation has been reported in a family of WD [7]. Blau syndrome is a paediatric granulomatous disease caused by NOD2 mutations of high penetrance in exon 4. The disease-associated NOD2 mutations precipitate a loss of canonical NOD2 signalling via RIP2 and decrease IRF4 expression [20]. The molecular mechanisms could partially explain the susceptibility of NOD2 mutations to TW or WD. It is worth mentioning that case 2 also carried a heterozygous MEFV variant nondiagnostic of classic FMF, a recessively inherited disease. Of note, MEFV mutations are resistant to certain bacteria like Yersinia pestis [21].

Figure 2.

Schematic representation of NOD2 protein and signalling pathway. NOD2 as intracellular sensor recognizes and binds bacterial peptidoglycans or MDP of bacteria via the LRR region and activates the NOD2 signalling pathway, leading to proinflammatory cytokine production and subsequent inflammatory response. NOD2: nucleotide binding oligomerization domain 2; MDP: muramyl dipeptide; LRR: leucine-rich repeat; TLR: Toll-like receptor; RIP2: receptor interacting protein kinase 2; P38/ERK/JNK: mitogen-activated protein kinase; NF-κB: nuclear factor-κB

Open in new tabDownload slide

Management strategy for WD patients with carriage of NOD2 mutations

Both WD and YAOS predominantly affect the Caucasian population (North America and Europe) perhaps in part because NOD2 LRR mutations are rarely detected in Asian and African populations [22, 23]. YAOS affects more females than males in a ratio of two to one, whereas WD primarily affects middle-aged men (85%). One explanation for the gender discrepancy and male predominance in WD is that androgen reduces peripheral blood monocytes and binds its receptors on macrophages to suppress proinflammatory cytokines and intestinal inflammation, and thus may further blunt clearance of pathogens like TW [24].

In addition to gastrointestinal symptoms, 85% of patients with WD have arthralgias (only 25% experience myalgias), while 45% report fevers and 40–50% develop pleural effusions [10]. Cutaneous presentation is rare, though erythema nodosum-like lesions are reported in antibiotic treated patients. YAOS patients share a similar clinical phenotype, but dermatitis that commonly presents as patchy erythema (75%) and eyelid swelling/discoloration are characteristic of YAOS rather than WD. These manifestations may be used to distinguish these two diseases. Ultimately, bidirectional endoscopy with tissue biopsies is key. Our data may also indicate the contributory role of autoinflammation to the long and insidious symptoms preceding the diagnosis of WD in these cases. Molecular analysis for NOD2 mutations in WD patients may provide much-needed therapeutic targets beyond the standard long-term and recurrent treatment with antibiotics, as up to 33% of patients relapse causing significant morbidity and mortality [10].

Potential treatment regimens for WD include short-term ceftriaxone and long-term trimethoprim–sulfamethoxazole or doxycycline and HCQ. In the described cases, all patients were on ceftriaxone then trimethoprim–sulfamethoxazole or doxycycline and HCQ. Their symptoms greatly improved with antibiotics but flares still occurred. This contrasts with antibiotic efficacy that has been described. For example, in a prior study of 13 patients, no relapses were observed after treatment with doxycycline and HCQ followed by doxycycline alone [25]. We hypothesize that WD patients with relapses or residual symptoms could have uncontrolled autoinflammation related to NOD2 mutations. The anti-inflammatory actions of both doxycycline and/or HCQ in certain rheumatic diseases may be applicable. In our study, case one received intermittent dosing of anakinra, and case two received one dose of canakinumab (with concurrent antibiotics) with prompt symptomatic improvement, suggesting that IL-1 inhibitors could be used intermittently (or with a longer interval considering potential immunosuppression) to better control residual or recurrent symptoms.

In conclusion, our case series indicates that NOD2 LRR-associated mutations in monocytes/macrophages may further impair the function of these cells and thus increase the host susceptibility to TW infection and development of WD, particularly in males exposed to immunosuppressive agents. A genetic analysis for NOD2 mutations using whole NOD2 gene sequencing may benefit WD patients in diagnosis and management, particularly those with relapses despite adequate antibiotics. Our report provides new insights into the pathogenesis of WD and patient care. Unlike common diseases that generally have guidelines and consensus, these are rare diseases as in our study, and best evidence may come from case reports and case series.

Data availability

Data are available upon reasonable request.

Contribution statement

Q.Y. and J.M.D. co-conceived and co-designed the study. Q.Y. wrote the first draft of the article. All authors provided clinical care to one or more patients and contributed to data collection. K.A.W. collected most data on these patients, participated in discussion and interpretation of the report, and critically formatted the manuscript. All authors participated in manuscript revision, agreed to submit the manuscript and approved the final version of the manuscript. Q.Y. and J.M.D. accessed and verified the data, and share the corresponding author role.

Funding

No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.

Disclosure statement: J.M.D. has a research grant from Pfizer, licenced technology to Girihlet and submitted a provisional patent application for assessing rheumatoid arthritis; all are unrelated to the present work. All other authors declare no conflicts of interest.

References