Introduction

Non-tuberculous mycobacteria (NTM) are ubiquitous, environmental organisms that are generally considered poorly pathogenic. Given this weak pathogenicity, predisposing factors for disease include older age, underlying lung disease, and immune compromise; disseminated non-tuberculous mycobacterial disease (dNTM) has historically been seen in the context of advanced Human Immunodeficiency Virus (HIV) [1–3]. Interferon-gamma (IFN-γ) autoantibody syndrome is an emerging clinical entity involving the presence of antibodies that neutralise the function of endogenous IFN-γ and is being linked with the development of dNTM disease in healthy young people, a population not previously thought to be at risk. We discuss the case of a 29-year-old Filipino male with disseminated mycobacterial infection who was found to have acquired IFN-γ autoantibodies.

Case presentation

A 29-year-old Filipino man was admitted from a visiting international cargo ship with 3 weeks of worsening fever, productive cough, and right-sided submandibular and axillary swelling. He had started feeling unwell six weeks prior with gradual onset low back pain, soaking night sweats and weight loss. He had no significant background history, regular medications or known allergies. He worked on an international cargo ship, residing in the Philippines between voyages, and denied alcohol, smoking, or illicit substance use. On examination he was cachectic, febrile, and had significant palpable lymphadenopathy in bilateral submandibular, cervical chain and axillary regions. There was no hepatosplenomegaly nor respiratory abnormalities. Computed tomography (CT) scans confirmed marked cervical, mediastinal, hilar and subdiaphragmatic lymphadenopathy as well as a lingular mass lesion (Fig. 1). Other initial investigations are summarised in Table 1; notably blood and urine cultures, multiple sputum samples (including examination for acid-fast bacilli and Mycobacterium tuberculosis (TB) Polymerase Chain Reaction (PCR)), and HIV, hepatitis B, hepatitis C and autoimmune serology were negative. TB IFN-γ release assay showed inadequate mitogen response repeatedly.

View Full Size

A) (left) showing axial sections of CT neck and B) (right) demonstrating axial images of CT chest. Multiple enlarged lymph nodes, including some with central necrosis (arrows), are seen in the submandibular, submental, and mediastinal regions

Citation: European Journal of Microbiology and Immunology 15, 1; 10.1556/1886.2024.00123

• Download Figure
• Download figure as PowerPoint slide

The patient deteriorated rapidly with a sepsis-like syndrome whilst trying to establish a diagnosis despite broad-spectrum antimicrobials including piperacillin-tazobactam. In addition, he developed significant pulmonary and peripheral oedema, weight loss, worsening anaemia and liver function test (LFT) derangement. A fluorodeoxyglucose positron emission tomography (FDG-PET) scan showed numerous hypermetabolic nodal foci in the neck, supraclavicular fossae, axillae, lungs, and the liver, and diffuse increased uptake throughout the spleen, most consistent radiologically with lymphoma (Fig. 2). A core lymph node biopsy was obtained on day 9 of admission but was non-diagnostic, only showing reactive lymphadenitis, necessitating an excisional lymph node biopsy on day 15. Shortly afterwards, multiple samples grew mycobacteria: the sputum samples after 14 days, and the initial lymph node biopsy after 5 days. These were identified as Mycobacterium persicum, a slow-growing NTM, and Mycobacterium abscessus, a rapid-growing NTM, respectively.

View Full Size

Coronal view of FDG-PET showing increased avidity throughout the lungs, liver, spleen, and multiple lymph nodes

Citation: European Journal of Microbiology and Immunology 15, 1; 10.1556/1886.2024.00123

• Download Figure
• Download figure as PowerPoint slide

He was commenced on empirical therapy for dNTM with amikacin, cefoxitin, rifabutin and azithromycin. Treatment was complicated by the development of severe dyspnoea and chest pain secondary to a large, exudative pericardial effusion close to tamponade; no acid-fast bacilli or other organisms were seen on microscopy, nor was there any growth, however a course of tapering prednisone was commenced to cover the possibility of an Immune Reconstitution Inflammatory Syndrome-like pathology. The diagnosis of dNTM was thought unusual in an immunocompetent host and a predisposing factor was sought. On review of the literature the syndrome had been described in association with antibodies inhibiting interferon-gamma, and subsequent immunological testing was consistent with the presence of these antibodies. They were detected via a functional assay: exogenous IFN-γ was added to our patient's sample, and any resulting STAT1 phosphorylation, a downstream effect of the IFN-γ pathway activation, was measured by fluorescence. In our case, phosphorylation was absent, indicating that blockade of the exogenous IFN-γ was occurring. He received a dose of rituximab intended to lower the titre of antibodies. Over the next two weeks the patient's clinical and laboratory parameters improved significantly and he was transferred to a tertiary hospital in the Philippines to continue treatment.

Discussion

Disseminated NTM disease has historically been associated with significantly immunocompromised hosts, particularly those with advanced HIV [1–3]. The association of dNTM disease in younger adults of South-East Asian descent with acquired autoantibodies to interferon-gamma (IFN-γ) is a relatively recent development and not widely recognised [4–8]. In the 1990s, Mendelian susceptibility to mycobacterial disease (MSMD) was discovered, with the mutations involved all affecting the function of the interleukin 12 (IL-12)-interferon gamma (IFN-γ) pathway [5]; this pathway has been found integral to the immune system's defence against mycobacteria [1]. In a functioning immune system, phagocytosis of a mycobacterium induces production of IL-12, which binds to and activates T-cells and Natural Killer cells, resulting in production of IFN-γ [1]. IFN-γ leads to a signalling cascade and upregulation of IFN-γ-responsive genes inducing further production of cytokines and macrophage activation, resulting in efficient containment of the intracellular mycobacteria by phagocytosis and/or granuloma formation [1].

In 2004, Doffinger et al. described a case of a young, previously healthy Filipino man without HIV who developed multiple disseminated mycobacterial infections. He was found to have defective IFN-γ production, and because this patient later developed multiple autoimmune conditions and did not respond to exogenous IFN-γ administration, the possibility of autoantibodies neutralizing cytokines or receptors in the IL-12/IFN-γ pathway was raised as the mechanism. A high titre of neutralising IFN-γ antibodies were subsequently found in the patient's serum [4]. Throughout the 2000s, further cases of previously healthy adults with dNTM in association with IFN-γ autoantibodies were described [6–8]. In 2007 Chetchotisakd et al. reported a case series of 129 patients in Thailand with dNTM without personal or family history of immunodeficiency, again indicating a possible acquired immunodeficiency, and later random sampling of the cohort found IFN-γ autoantibodies in these patients [6]. Of note, dNTM with acquired IFN-γ autoantibodies is most common in people of East Asian descent, indicating a possible genetic link, however they have also been seen in a wide range of ethnicities including British, German and South African [2, 8].

Initially the role of the IFN-γ autoantibodies in these patients was unclear as they had also been found in patients without active infection [9]. In 2012, Browne et al. [9] conducted a study on Taiwanese and Thai adults without previously known immunodeficiency who had opportunistic infection either with dNTM alone, other opportunistic infection, or a combination of the two, and compared them with tuberculosis patients and healthy controls. They found that neutralising IFN-γ autoantibodies distinguished the groups with opportunistic infections from the TB and healthy control patients; 2% of both the pulmonary tuberculosis and healthy controls had IFN-γ autoantibodies compared with >80% for the groups with opportunistic infection including dNTM, indicating that mycobacterial infection itself was not the catalyst for the presence of the antibodies [9]. In 2013 Wongkulab et al. [10] found a similar, significant association between the presence of IFN-γ autoantibodies and adult-onset immunodeficiency syndrome, and that there was a correlation between higher antibody titre and active infection i.e. those who didn't have infection tended to have lower antibody levels. Yoshizawa et al. (2020) have also found that the IFN-γ titre fluctuated as the severity of the dNTM disease fluctuated but did remain at a detectable level even with improvement on treatment [11].

The detection of neutralising IFN-γ autoantibodies can be achieved in several ways. IFN-γ release assays such as QuantiFERON-TB can indicate the presence of the antibodies [12]. One component of the test is the mitogen or positive control which measures the response to a non-specific T-cell stimulant [12]. It has been demonstrated that patients with detectable neutralising IFN-γ autoantibodies have a consistently insufficient mitogen response resulting in an invalid QuantiFERON result, with the advantages of this method being the availability and relatively rapid turn-around time [12]. Targeted assays for detecting IFN-γ autoantibodies are generally only available in-house in Immunology reference laboratories and include functional assays, Western blot, and Enzyme-Linked Immunosorbent Assays (ELISA) [12].

Anti-cytokine autoantibodies and their role in disease remains a growing area of research. A range of other anti-cytokine autoantibodies have been discovered in recent decades in association with various infections, rheumatological conditions and malignancies, such as Granulocyte colony-stimulating factor (G-CSF) autoantibodies in Felty's syndrome, Granulocyte-macrophage colony-stimulating factor (GM-CSF) autoantibodies in cryptococcosis, and interleukin 6 (IL-6) autoantibodies in recurrent staphylococcal skin infections [13–15]. It is thought that as the knowledge base increases, other conditions previously described as idiopathic may become associated with anti-cytokine autoantibodies [9]. Their role in pathogenesis also provides important avenues for treatment options such as therapeutic antibodies and biologic agents aimed at the autoantibodies [12].

The next consideration is whether the presence of IFN-γ autoantibodies has any other implications for patients in terms of prognosis, treatment, and future sequelae. Regarding dNTM infection itself, in general immunocompromised patients have poorer outcomes than immunocompetent patients: in 2022 Chai et al. [3] found 5-year mortality rates at 12.9% and 35.9% for immunocompetent and immunocompromised patients respectively. Additionally, the diagnosis of the dNTM in IFN-γ patients is often delayed. In one case series the average time to diagnosis after symptom onset was eight months [2], which may affect outcomes. Given the defective cell-mediated immunity, patients with IFN-γ autoantibodies can develop infections with other organisms, particularly Salmonella spp., cytomegalovirus, varicella-zoster virus, Cryptococcus spp., Talaromyces marneffei, Histoplasma capsulatum and Aspergillus spp., as well as reactive dermatoses such as Sweet syndrome [6, 7, 15]. In a 2018 longitudinal cohort study of adults who were diagnosed with IFN-γ autoantibodies in Thailand and the United States (most of whom were of Southeast Asian descent), 64% had presented with dNTM disease, 23% with another opportunistic infection plus dNTM, and 14% with another opportunistic infection alone. Among these opportunistic infections, Salmonella spp. and varicella-zoster virus were the most common [16].

Another significant factor impacting outcomes is that given dNTM with IFN-γ autoantibodies is a relatively new and rare clinical syndrome, there is uncertainty regarding optimal management [13]. In the same 2018 longitudinal cohort study, the Thai patients were mostly treated with antimicrobial agents alone without secondary prophylaxis after treatment of the index infection. A larger proportion of the US cohort (39%) received immunomodulatory therapy with rituximab in addition to antimicrobials, as well as all being continued on indefinite secondary prophylaxis with azithromycin after treatment of the index infection. 24% of the Thai patients died during follow up and mean percentage of follow up visits with ongoing infection was 29%, compared with no deaths in the US cohort and 14% with ongoing infection, although it should be noted that a higher proportion of Thai patients were infected with M. abscessus, a rapid growing NTM [16]. It is generally accepted that patients should receive standard combination anti-mycobacterial therapy empirically, though treatment durations have ranged from 18 months to seven years, and it is as yet unclear whether secondary prophylaxis is beneficial after completion of primary treatment [2]. Regarding treatment of the circulating autoantibodies, rituximab is the most evidence-based immunological therapy so far; other agents trialled include cyclophosphamide, intravenous immunoglobulin, and Rituximab-Cyclophosphomide-Doxorubicin-Vincristine (R-CHOP) chemotherapy. Rituximab and cyclophosphamide have been found to significantly lower IFN-γ autoantibody titre [16]. Immunomodulatory therapy has generally been reserved for severe dNTM disease despite initiation of antimicrobials, and time to commencement has ranged from three months to more than a year [2]. Our patient received a second dose of rituximab a month later and had approximately 18 months of combination NTM treatment. He has remained well without further hospital admissions six months after cessation.

Conclusion

There is an increasing number of reports of dNTM infection in previously healthy younger adults, particularly in those of South Asian descent, and a strong association with acquired neutralising autoantibodies to IFN-γ has been found. dNTM should be considered in all patients presenting with similar clinical syndromes, not just those with pre-existing immunosuppression, and if diagnosed, evaluation for IFN-γ autoantibodies should occur. Further research is required particularly regarding long-term outcomes to better inform prognosis and optimal care.

Consent

Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.

Funding sources

No financial support was received for this study.

CRediT author statement

Victoria Jordan: Writing – original draft. Robert Pickles: Writing – Review & Editing, Supervision.

Conflict of interest

The authors declare that there is no conflict of interest regarding the publication of this article.