Fatal Systemic Granulomatous Disease Associated with Vaccine-Derived Rubella Virus in AIOLOS-Associated Immunodeficiency

Abstract

We present a case of disseminated granulomatous disease associated with live attenuated vaccine-derived rubella virus (VDRV) in an immunodeficient patient who subsequently developed fatal multi-organ failure, including neurological deficits, with VDRV detected in the cerebrospinal fluid. Posthumous genetic analysis found a heterozygous missense mutation in IKZF3, a gene where haploinsufficiency is linked to immunodeficiency and immune dysregulation. Functional studies demonstrated that this mutation decreases AIOLOS protein stability and half-life, and family members carrying the same mutation exhibited decreased AIOLOS levels. This report underscores the importance of considering persistent rubella virus (RuV) infection in patients with cutaneous granulomatous lesions and highlights the need for comprehensive evaluation to uncover potential underlying immunodeficiencies, which could inform and optimize individualized treatment strategies.

Introduction

Granulomatous disease caused by vaccine-derived rubella virus (VDRV) has predominantly been reported in patients with primary immune deficiencies [7]. Further reports have expanded the landscape of susceptibility to include common variable immunodeficiency and isolated immunoglobulin deficiencies in otherwise immunocompetent individuals [1], [5]. Reports of granulomas associated with wild type RuV arising in older individuals show that RuV infections contracted anytime in life can persist long term and re-emerge decades later [8].

Reported treatments for RuV-associated granulomas have included oral and topical steroids, broad-spectrum antivirals such as ribarivin, and TNF-alpha and IL-1R inhibitors, with limited success. Immune reconstitution after hematopoietic stem cell transplantation in immunodeficient patients presenting with VDRV have shown to effectively control this complication, albeit in a small number of cases. A critical question remains whether RuV-associated granulomas should be managed with immunosuppression or immune activation, both of which have reported efficacy.

Mutations in the Ikaros Zinc Finger Transcription Factor (IKZF) family, including AIOLOS (IKZF3) and IKAROS (IKZF1) are increasingly linked to immunodeficiency disorders. These transcription factors work as homo- or heterodimers, frequently down-regulating the activity of their target genes, primarily in pericentric-heterochromatin areas. Germline AIOLOS mutations acting by dominant negative (DN) mechanisms are associated with as B-cell deficiency, abnormal T-cell profiles, and increased susceptibility to different types of infections and malignances [3, 9, 10]. In addition to the AIOLOS DN variants, patients with AIOLOS haploinsufficiency can present with hypogammaglobulinemia, recurrent infections, autoimmunity, and allergy, with incomplete clinical penetrance [4, 6]. Further characterization is needed to identify affected individuals and guide targeted interventions.

In this report we describe a case of fatal disseminated VDRV infection in an immunodeficient patient who presented with extensive cutaneous granulomas and severe multi-organ disease, including neurological deficits. Posthumous genetic analysis revealed an IKZF3 missense mutation associated with decreased AIOLOS protein stability, which to our knowledge, has not previously been linked to RuV-associated granulomas. We believe that the presence of RuV should be considered in patients with suggestive cutaneous granulomatous lesions, exploration for underlying immunodeficiency may be warranted, and immunosuppression should be employed cautiously, particularly for infection-associated granulomas associated with immunodeficiency.

Materials and Methods

Sanger Sequencing

For Sanger sequencing of IKZF3 (NM_012481.5): c.229A > T, p.M77L, genomic DNA was PCR-amplified using GoTaq DNA Polymerase (Promega) and variant-specific, M13-tailed oligonucleotide primers as follows: forward: 5’-agtatggcttcgcttatgaacca-3’ and reverse: 5’-ttcttgcattttcttccctgcag-3’. PCR products were purified using ExoSAP-IT reagent (Thermo Fisher Scientific) and were directly sequenced using BigDye Terminators (version 1.1) and universal M13 forward and reverse primers. Sequencing products were analyzed using a 3500xL Genetic Analyzer (Thermo Fisher Scientific). For sequencing from cDNA, total RNA was isolated from PBMCs with the RNeasy plus mini Kit (QIAGEN) and the RNA was reverse transcribed using QuantiTect Reverse Transcription Kit (QIAGEN). cDNA was amplified by PCR using GoTaq DNA polymerase with the following primers: forward: 5’- cagcccggcccggcagcgac-3’ and reverse: 5’-cagaatgtgtcctaagatgc-3’.

Next-generation sequencing

Whole exome sequencing (WES) was performed using genomic DNA with the Illumina Exome with Enrichment kit and NextSeq 2000 instrument according to the manufacturer’s protocols. Reads were mapped to the UCSC Genome Browser hg38 assembly and variants were called using the Illumina DRAGEN Enrichment Pipeline. Variants were annotated using ANNOVAR (https://annovar.openbioinformatics.org/en/latest/misc/credit/).

Immunofluorescence Staining

NIH3T3 cells were transfected with pcDNA3-HA-AIOLOS WT or M77L using the Nucleofector kit R (Amaxa, program A-24) according to the manufacturer’s instructions. Transfected cells were seeded onto coverslips and incubated for 24 h. The cells were washed with PBS and then fixed for 10 min in 4% paraformaldehyde and permeabilized for 15 min in 0.1% Triton X-100 in PBS at room temperature. After permeabilization, the cells were blocked in blocking buffer (PBS with 10% FBS and 0.1% Triton X-100) for 30 min and then incubated for 2 h with an anti-HA antibody (Biolegend, Cat. 901,501), followed by an Alexa Fluor 488-conjugated secondary antibody (Thermo Fisher Scientific, Cat. A11001) for 1 h. Cells were washed with PBS two to three times between each step. Cells were mounted on slides using VECTASHIELD Mounting Medium (Vector Laboratories) and visualized using an EVOS M5000 cell imaging system (40X objective, Thermo Fisher Scientific).

RuV RT-PCR, Serology, and Immunohistochemistry

Methods for RuV testing have been described previously [11]. Briefly, Rubella IgG titers were measured by an enzyme-linked immunosorbent assay (EIA) with the ZEUS ELISA Rubella IgG Test System (ZEUS Scientific, Branchburg, NJ) according to the manufacturer’s instructions. For immunohistochemistry, FFPE Sects. (3–4 µm) were deparaffinized, rehydrated, and retrieved, then washed, permeabilized, and blocked. Sections were incubated overnight (4 °C) with mouse anti–rubella capsid mAb (± rabbit cell markers), followed by fluorescent secondary antibodies. Primers and conditions for real-time RT-PCR for RuV, and a detailed strategy for a sequencing have been described [7].

Cycloheximide Chasing Assay

HEK293T cells were transfected with pcDNA3-HA-AIOLOS WT or M77L using Effectene (QIAGEN) according to the manufacturer’s instructions. After 16–18 h of incubation, cells were treated with cycloheximide (10 μg/ml, Sigma-Aldrich Cat. C4859) for additional 24 h. Cells were lysed in the lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 2 mM EDTA, 0.5% Triton X-100 and 0.5% NP40, and protease and phosphatase inhibitor cocktail (Sigma, PPC1010) and the protein lysates were subjected to the immunoblotting using anti-HA (Cell signaling technology, Cat. 2999S) and anti-Vinculin antibodies (Santa Cruz Biotechnology, Cat. sc-73614). The images were acquired and analyzed using a C-Digit scanner and Image Studio software (LICORbio), respectively.

Immunoblotting

CD3 T cells were enriched from total PBMCs using an EasySep Human T cell enrichment kit (STEMCELL Cat. 19051) according to the manufacturer’s instructions. Enriched CD3 T cells were washed with PBS once and lysed in the RIPA buffer (Thermo Fisher Scientific, Cat. 89900) including the protease and phosphatase inhibitor cocktail (Sigma, PPC1010). Cell lysates were subjected to immunoblotting using anti-AIOLOS (Cell Signaling Technology, Cat. 15103S) and β-actin (Santa Cruz Biotechnology, sc-47778) antibodies. The images were acquired and analyzed using an iBright imaging system and iBright analysis software (Thermo Fisher Scientific), respectively.

AIOLOS Protein Expression by Flow Cytometry

AIOLOS expression in T and B cells were measured as described previously [4].

Patients and Samples

All patients or their guardians provided written informed consent in accordance with the Declaration of Helsinki under institutional review board − approved protocols. Blood from healthy donors was obtained under approved NIH protocols.

Results

Clinical Course

The patient was a previously healthy man who first presented in 2007 at age 26 after an occupational exposure with wet cement. He subsequently developed irritated, erythematous skin lesions at the exposure site, initially diagnosed as third-degree burns. Despite debridement and wound care, the lesion persisted, and a skin biopsy demonstrated granulomatous inflammation. Initial treatment with etanercept was unsuccessful.

The patient progressively developed widespread, similarly granulomatous patches and plaques on his arms, legs, and face. Workup in 2009 revealed leukopenia, lymphopenia, and hypogammaglobulinemia (IgA: 13 mg/dL, IgG: 158 mg/dL, IgM: 33 mg/dL), leading to the presumptive diagnosis of common variable immunodeficiency (CVID), the most prevalent form of adult-onset primary immune deficiency/inborn error of immunity (IEI). He was managed with subcutaneous immunoglobulin (IVIG) therapy, which stabilized his skin condition for several years.

In 2013, the patient re-presented with extensive skin lesions (Fig. 1A). Granulomatous inflammation was re-demonstrated on skin biopsy (Fig. 1B), and all four skin biopsies taken both at this time and four years later revealed the presence of RuV capsid protein in M2 macrophages, confirmed via immunohistochemistry (Fig. 1C). Over the next several years, he received a variety of immunomodulatory and immunosuppressive therapies, including hydroxychloroquine, doxycycline, systemic corticosteroids, methotrexate, and infliximab. His skin lesions responded well to this combination, with complete resolution of his granulomatous dermatitis while on infliximab. However, his disease course was complicated by the development of lower extremity ulcers and opportunistic infections, including Candida spp. esophagitis, herpes simplex virus esophagitis, a retropharyngeal abscess, extensive human papillomavirus associated verruca, and Listeria spp. bacteremia, prompting periodic suspension of immunosuppressive treatment.

Fig. 1.

Clinical features in a patient with CVID andRuV-associated granulomatous dermatitis. A. Photos showing widespread granulomatous skin lesions at initial clinical presentation. B. H&E stained images of the patient’s skin biopsy, which demonstrate granulomatous dermatitis with multinucleated giant cells. C. Immunohistochemistry for RuV capsid protein (red), CD206, a marker of M2 macrophages (green), and cell nuclei (blue) performed on biopsy from lesion on patient’s left arm. D. T2 FLAIR brain MRI showing CNS lesions corresponding to symptoms of hearing loss. Yellow arrows demonstrate cranial nerves enhancement in bilateral internal auditory canal and pachymeningeal enhancement in bilateral frontal convexity. E. Lower extremity ulcers present in patients final hospital stay.

In 2019, after approximately one year off infliximab but while still on methotrexate, prednisone, and hydroxychloroquine, the patient had developed joint pain, abnormal gait, balance issues, and progressive hearing loss. A brain MRI revealed pachymeningeal enhancement along the trigeminal nerves, lower cranial nerves, cavernous sinuses, and internal auditory canal (Fig. 1D). Following a negative infectious workup, these lesions were ultimately thought to be due to CNS involvement of his granulomatous disease. Despite a period of neurological stabilization after reintroduction of infliximab, his condition deteriorated three years later, with worsening inflammatory arthritis, progressive loss of vision, taste, and hearing, anosmia, and new skin lesions, including lower extremity ulcers (Fig. 1E). He also developed pancolitis, numerous liver lesions, and a left basal ganglia stroke.

Molecular testing (reverse-transcription polymerase chain reaction (RT-PCR) and sequencing) of nasopharyngeal swabs confirmed the presence of VDRV infection. Metagenomic next-generation sequencing showed VDRV presence in the patient’s cerebrospinal fluid (CSF). RuV neutralizing antibody titer was elevated in the patient’s serum (1280 U/ml), which is typical for granuloma patients with RuV-associated granulomas [11]. Additional infectious workup was negative. Despite aggressive immunomodulatory therapy, antiviral treatments (ribavirin and nitazoxanide), and the addition of tofacitinib as an anti-granulomatous agent, the patient unfortunately succumbed to multiorgan failure in 2022.

Genetic and Molecular findings

Posthumous genetic testing via whole exome sequencing revealed that the patient (henceforth referred to as II.1), harbored a heterozygous missense variant in IKZF3, resulting in a methionine to leucine change: chr17(GRCh38): g.39792868T > A, NM_012481 (IKZF3), c.229A > T, p.M77L (Fig. 2A). This variant (NCBI dnSNP ID rs369652538) is reported in gnomAD v4.1.0 at a frequency of 0.00006816, with no homozygous occurrences reported in 1,613,904 alleles analyzed. The CADD GRCh38-v1.6 phred score is 10.82, and the variant is classified as “uncertain significance” according to American College of Medical Genetics and Genomics (ACMG) guidelines (PM2, PP5, BP4). IKZF3 mutations are known to give rise to immunodeficiency and immune dysregulation via haploinsufficiency [4]. No other variants in genes associated with primary immunodeficiency syndromes were detected.

Fig. 2.

Genetic and functional studies of the heterozygous M77L mutation in IKZF3 A. A schematic of AIOLOS zinc finger (ZF) domains and the M77L mutation is shown. B. The pedigree of a family with the heterozygous IKZF3 variant is shown, with black representing symptomatic carriers and gray representing an asymptomatic mutation carrier. A diagonal line across the symbol indicates a deceased individual, and the arrow indicates the proband. C. The chromatograms show the sequence of the heterozygous mutation (c.229A > T) in cDNA prepared from PBMCs. D. AIOLOS protein expression in CD3 T cell lysates from the mutation carriers and healthy controls (including II.3, a mutation-negative family member among the controls). Densitometry analysis of AIOLOS protein expression normalized by a loading control. The average expression value of five healthy controls was set to 100. Open circle indicates the mutation negative family member control (II.3). Data are presented as means ± SD. E. Intracellular AIOLOS expression in CD45RA⁺ (CD45RA⁺CD45RO⁻) and CD45RO⁺ (CD45RA⁻CD45RO⁺) T cells, and in naïve (IgD⁺CD27⁻) and memory (IgD⁻CD27⁺) B cells. AIOLOS MFI values were normalized to the average values of the corresponding HC naïve subset (CD45RA⁺CD45RO⁻ for T cells, IgD⁺CD27⁻ for B cells) and relative protein expression is displayed. Data represent means ± SD from three experiments (HCs; each dot indicates an individual HC), I.1 (n = 2 replicates), and II.2 and II.3 (n = 3 replicates). F. NIH3T3 cells were transfected with HA-tagged AIOLOS WT or the mutant. The cells were labeled with an anti-mouse HA, followed by an Alexa 488-conjugated secondary antibody. Cells were visualized using an EVOS (40X objective) fluorescent microscope. Scale bars indicate 25 μm. Data shown are representative of three independent experiments. G. HEK293T cells were transfected with HA-tagged AIOLOS WT or the mutant. The next day, cells were treated with cycloheximide (CHX, 10 μg/ml) for 24 h. AIOLOS expression was normalized to a loading control, and AIOLOS protein stability was calculated by dividing the CHX-treated sample by the untreated sample (X100) for each group. Data are presented as means ± SD. Data shown are representative of three independent experiments. Significance was determined by two-tailed Student t test, * p < 0.05

Whole exome and Sanger sequencing of the patient’s family members confirmed that the patient’s mother (I.1) and one brother (II.2) carried the same variant as the patient (II.1), while another brother (II.3) did not (Fig. 2B). Detection of this variant in cDNA from peripheral blood mononuclear cells (PBMC) of I.1 and II.2 indicated that it is transcribed and potentially expressed in PBMCs (Fig. 2C). Pertinent medical history for the mother (I.1) included psoriasis and lichen planus without reported history of serious infectious. The brother II.2 had no history of autoimmune disease or serious infections. Lymphocyte phenotyping in I.1 and II.2 were overall within normal limits, however the index patient (II.1) had very low B cells, reduced T-cell numbers, and borderline low NK-cell levels (Table 1).

Table 1.

Laboratory results of immunological tests

	I.1	II.1	II.2	Reference Ranges % (Abs number)
CD3 +	79.9% (1278)	74.8% (381)	68.5% (849)	55.3–88.7% (651–2804)
CD3 + CD4 +	42.7% (683)	40.8% (208)	37% (459)	27.9–55.8% (370–1336)
CD3 + CD8 +	26.2% (419)	26.5% (135)	28.5% (353)	13.6–46.2% (185–1024)
CD3 + CD4 + CD45RA + CD62L +  (Naïve cells, of CD4)	27.4%	n.d	32.3%	14–67%
CD3 + CD4 + CD45RA-CD62L +  (Central memory, of CD4)	43.8%	n.d	39.4%	26–64%
CD3 + CD4 + CD45RA-CD62L- (Effector memory, of CD4)	28.0%	n.d	27.3%	4.5–30%
CD3 + CD4 + CD45RA + CD62L- (TEMRA of CD4)	0.9%	n.d	1.0%	0–3.7%
CD3 + CD8 + CD45RA + CD62L +  (Naïve cells, of CD8)	41.6%	n.d	29.3%	25–73%
CD3 + CD8 + CD45RA-CD62L +  (Central memory, of CD8)	12.2%	n.d	11.1%	5.9–40%
CD3 + CD8 + CD45RA-CD62L- (Effector memory, of CD8)	24.0%	n.d	22.9%	5.5–34%
CD3 + CD8 + CD45RA + CD62L- (TEMRA, of CD8)	22.2%	n.d	36.7%	4.8–33%
CD4 + CD45RA + CD31 +  (RTE, of CD4)	15.8%	n.d	21.9%	6.7–45%
CD4 + CD45RA-CXCR5 +  (Tfh, of CD4)	15.1%	n.d	11.2%	3.7–14.1%
CD4 + CD25 + FOXP3 +  (Treg, % of CD4)	11.1%	n.d	5.9%	4–10%
CD16 + CD56 + (NK)	11.4% (182)	24.3 (124)	20.1% (249)	7.3–33.4% (126–841)
CD19 + (or CD20 +)	8.6% (138)	0.5% (3)	11.2% (139)	3.8–18.0% (79–399)
CD20 + CD27-IgM +  (Naïve B-cells, of CD20)	66.1%	n.d	87.6%	56–92%
CD20 + CD27 + IgM- (Switched memory, of CD20)	13.5%	n.d	6.1%	4.3–20%

Bold values indicate above the normal range, and bold and italics values indicate below the normal range; all based on NIH Clinical Center adult reference ranges. n.d. indicates not determined

AIOLOS protein expression was evaluated in CD3 T cells from available family members and healthy controls by immunoblotting. All individuals carrying the heterozygous IKZF3 M77L mutation showed AIOLOS protein levels approximately 30–50% lower than those in healthy controls, while the mutation-negative family member (II.3) showed levels comparable to the HCs (Fig. 2D). We also evaluated AIOLOS protein expression via flow cytometry in naïve (CD45RA⁺RO⁻ for T cells and IgD⁺CD27⁻ for B cells) and memory subsets (CD45RA⁻RO⁺ for T cells and IgD⁻CD27⁺ for B cells) of T and B cells from available family samples. AIOLOS expression was decreased in the mutation carriers, with reductions of approximately 25–33% in CD4⁺ T cells, 10–15% in CD8⁺ T cells, and a subtle decrease of 4–11% in CD19⁺ B cells compared to healthy controls (Fig. 2E).

To understand the functional impact of the M77L mutation on AIOLOS activity, recombinant constructs of HA-tagged wild type AIOLOS (AIO-WT) and AIOLOS with the M77L mutation (AIO-M77L) were generated and tested for their ability to target pericentromeric heterochromatin and for protein stability. AIO-M77L protein showed the punctate staining pattern characteristic of pericentromeric localization, suggesting that the mutant has normal pericentromeric heterochromatin targeting, similar to the WT control (Fig. 2F). We next tested the stability of the AIO-M77L protein. After treating the transfected cells with the protein synthesis inhibitor cycloheximide, we observed increased degradation of AIO-M77L compared to the AIO-WT control, with approximately 50% versus 11% reduction, respectively (Fig. 2G). These results suggest that the M77L mutation does not appear to affect the AIOLOS’ function, but may lead to protein instability, resulting in reduced AIOLOS protein levels.

Discussion

This case highlights the potential severity of RuV-associated granulomas and the challenges in selecting appropriate treatments, particularly in the context of immunodeficiency.

Critical questions remain regarding the optimal therapeutic approach—whether immunomodulatory, immune-stimulating, or antiviral therapies are most effective. While our patient initially clinically improved on a regimen of multiple immunomodulatory agents, most notably infliximab, and appeared to worsen with discontinuation, it is important to consider the potential role immunosuppression may have played in the widespread nature of his RuV infection.

We believe that our patient’s clinical course was likely influenced by his IKZF3 missense variant, which, based on data, leads to decreased levels of AIOLOS due to reduced protein stability. Recently, IKAROS variants in the early N-terminal region (pre-ZF1 region) have been shown to be associated with protein instability, leading to decreased protein expression [2]. Variants in this protein stability associated region (PSAR) do not appear to affect IKAROS function, as evidenced by normal peri-centromeric targeting and DNA binding. However, protein expression is decreased in patients’ primary cells due to decreased protein stability, resulting in haploinsufficiency and association with immunodeficiency. An AIOLOS variant (E82K) in the early N-terminal region also showed decreased protein expression in patients’ primary cells due to instability and is linked to hypogammaglobulinemia, recurrent infections, and a CVID-like phenotype [4], as observed in this patient, though how it may give rise to RuV-associated granulomas remains unclear. Future work understanding this relationship is crucial to identifying whether AIOLOS-deficient individuals are uniquely vulnerable to disseminated RuV granulomatous disease and to identify effective treatment strategies.

Overall, we propose that patients with disseminated cutaneous granulomatous inflammation, particularly with the histologic features historically described in RuV-associated granulomas (lymphocyte-rich granulomatous inflammation affecting the skin) should undergo RuV testing and comprehensive immunodeficiency evaluation, including for IKZF3 and other mutations associated with PID/IEI While immunosuppression may provide symptom relief, this case highlights concerns that this approach may exacerbate viral persistence and potentially contribute to dissemination and disease progression. If RuV is detected, RuV-specific antiviral and immune-stimulating therapies should be explored. For severe or refractory cases, hematopoietic stem cell transplantation remains the most reliable option.

Author Contributions:

Linda Zhou, Thomas H. Leung, Karolyn A.Wanat, Kathleen Sullivan, Dayna Gager, Matthew Helm, Galen Foulke, Misha Rosenbach: These authors participated in the clinical care and diagnostic course of this patient Hye Sun Kuehn Agustin A. Gil Silva, Ludmila M. Perelygina, LiJuan Hao, Min-hsin Chen, Julie E. Niemela, Jennifer L. Stoddard, Sergio D. Rosenzweig: These authors performed experimental work.

The contributions of the NIH authors are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.

Funding

This work was supported by the NIH Undiagnosed Disease Network, the Intramural Research Program, the NIH Clinical Center, and the National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH), US.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Consent to publish

The authors affirm that the family of the patient described provided informed consent for publication of the images in Fig. 1.

Conflict of interest

The authors declare no competing interests.