Exploring Genetic Clues in Kikuchi-Fujimoto Disease: A Three-Generation Study

Abstract

Background

Kikuchi-Fujimoto disease (KFD) is a rare disorder involving necrotizing lymphadenitis which was first described in Japan. The presentations include lymphadenopathy, fever, and leukopenia of unknown cause.

Patient

A 2-year-old Malay girl was evaluated and diagnosed with KFD after presenting with prolonged fever and rash. Her parents are not biologically related, but there is a strong family history of KFD on the maternal side of the family. We investigated the genetic profiles of the patient, her mother and grandmother, who also had a history of similar presentations.

Methods

High-resolution HLA typing was performed on the family to examine the potential associations with HLA genotypes. Next-generation sequencing using a Mendelian gene panel was performed on the child’s sample, followed by targeted Sanger sequencing of shortlisted variants for the family members.

Results

The three affected family members share 14 HLA alleles that are not present in the healthy father, among which five have reported associations with autoimmune inflammatory disorders. Two of them (DQA1*02:01 and DRB1*07:01) were also reported in a pair of affected twins from Australia. In addition to HLA, three rare missense variants in immune-related genes (FGFR3, NRP2 and PCM1) were found to correlate with the presence of symptoms within the family. The three affected members also had at least one of the variants in the C1QC gene (a missense and a 17-nucleotide deletion), which is part of the complement pathway that might be related to KFD.

Conclusion

Our study uncovers some HLA alleles and immune gene variants that might be important in the development of KFD. Large-scale sequencing of immune genes may be as important as high-resolution HLA typing in identifying genetic factors contributing to KFD pathogenesis. HLA studies provide risk associations to KFD, while association with variants in immune genes may reveal pathways in the development of KFD.

Introduction

Kikuchi-Fujimoto disease (KFD), also known as histiocytic necrotizing lymphadenitis, was first described in 1972 in Japan.¹ Patients commonly present with cervical lymphadenopathy which may be accompanied by fever and weight loss. The symptoms can manifest over several weeks and can be associated with infections, autoimmune conditions, or less commonly, malignancies. As the symptoms are not specific, the diagnosis of KFD is challenging and can be easily missed. A diagnosis can be confirmed by lymph node biopsies, which show histological features of necrotic foci with apoptotic debris, and accumulation of histiocytes, some with crescent-shaped nuclei. Another histopathological feature of KFD is the absence of neutrophils and eosinophils. Immunohistochemical staining will show myeloperoxidase and CD68-positive histiocytes. Immunoblasts, predominantly CD8+ T cells, are also present, while B cells are less frequently observed.² Histologically, KFD may appear very similar to systemic lupus erythematosus (SLE) lymphadenitis except for the absence of hematoxylin bodies that are found in SLE lymphadenitis. Other differential diagnoses include infectious lymphadenitis, autoimmune lymphadenopathy, and also malignancies.

KFD is a self-limiting disease which typically resolves spontaneously within 1–4 months, and patients usually only require symptomatic relief.³ The recurrence rate is low at 3–10%, but it can be up to 42.4% for childhood cases. Systemic corticosteroids are frequently used in relapsing cases or those with more severe symptoms.²

Cases of KFD were first described in Japanese followed by other Asian women, but have since been extended to include different ethnic backgrounds and both genders.¹ Most cases are diagnosed in young adults under the age of 40. Recent studies in Japan and Korea found a higher incidence in males for patients under the age of 20 years, but higher in females for the above 20-year-old age group.^3,4 The recurrence rate also follows a similar pattern.

Although it has been 50 years since KFD was first described, the etiology and pathogenesis remain unclear. One hypothesis is that it is triggered by infection with viruses such as Epstein-Barr virus (EBV) and herpes viruses.² However, the link between any virus and KFD has not been convincingly demonstrated. Another theory is that KFD is a result of autoimmune dysfunction, as KFD has histological features that appear similar to other autoimmune disorders, and symptoms matching KFD have also been reported in patients with other autoimmune disorders.⁵ In particular, SLE shares many characteristics with KFD and some patients had been diagnosed with both concurrently or one after the other.⁶

Previously reported cases of affected siblings and parent-child pairs suggest a genetic component in the development of the disease.^7–11 A number of studies on HLA had been done. In Japanese patients with KFD, frequencies of DPA1*01 and DPB1*02:02 were higher than the general population.¹² However, these do not correspond to HLA reported in patients from other countries. Other genetic studies have also identified loss-of-function mutations in immune regulating genes including complement pathway genes; and MEFV gene.^13–17 Although there were only a few such studies, it shows the potential of next-generation sequencing in advancing the understanding of this rare disease.

In this report, we present a familial case of KFD and investigate the genotypes of several candidate genes in immediate family members. Written informed consent for genetic investigation and publication from her parents was obtained under a research study approved by the SingHealth Centralised Institutional Review Board (CIRB 2014/571/f).

Case Presentation

Patient Info

A female child of Malay ancestry presented at 2 years and 1 month of age, with 3 weeks of fever, lymphadenopathy and 2 weeks of rash. She had been treated 2 weeks prior for an E. coli urinary tract infection which had cleared on subsequent testing.

The parents were not biologically related. However, there is a strong family history of similar presentations on the maternal side. Her mother, maternal grandmother, maternal uncle and his daughter all presented with similar episodes of prolonged fever or rashes when they were younger. The maternal grandmother was hospitalized for fever and rash. There was a record of lymph node biopsy but the result could not be retrieved. Her mother had similar presentations but did not seek treatment at a medical facility. The uncle’s daughter had a lymph node biopsy and was diagnosed as KFD based on histopathological examination. The uncle had prolonged fever for a few weeks during his childhood and was eventually diagnosed as having extranodal KFD.

Clinical Findings and Diagnosis

The rashes developed two weeks after the start of the fever. It initially involved the arms but subsequently progressed to the trunk, face, and lower limbs (Figure 1). It had an initial appearance of erythematous macules and papules that evolved into targetoid lesions with central areas of duskiness and limited areas of tense bullae. There was no mucosal involvement, no conjunctivitis, and no genital ulcers. Nikolsky’s sign was negative. When the fever did not resolve and lymphadenitis developed, she was given a presumptive diagnosis of KFD. There were multiple subcentimeter cervical lymph nodes bilaterally, with the largest lymph node on the right measuring about 1 cm in diameter. She had no organomegaly, arthritis or other lymphadenopathy. No facial or joint swelling was observed. There were no other systemic presentations such as loss of weight or appetite, night sweats, joint pains, or ulcers. The patient had sustained a cut one week before, but had no significant contact, exposure or travel history.

Figure 1.

Rashes seen on the patient’s (A) face and (B) body during initial presentation.

The patient’s blood test done at admission showed mild leukopenia, elevated inflammatory markers, including C-reactive protein (CRP) at 11.2 mg/L (normal 0–5 mg/L), procalcitonin 0.43 µg/L (normal < 0.09 µg/L) and erythrocyte sedimentation rate (ESR) at 25 mm/hr (normal 3–15 mm/hr), and mild transaminitis (Table 1). Serology results for cytomegalovirus, EBV, mycoplasma, toxoplasma, anti-streptolysin O titre (ASOT) and dengue were negative. Blood and urine cultures were also negative, as were autoimmune markers, complements, anti-nuclear antibody and anti-double-stranded DNA antibodies. An ultrasound of the neck confirmed several prominent bilateral cervical lymph nodes, with the most prominent lymph nodes at the right anterior neck abutting the right submandibular gland measuring 1.9 cm × 1.0 cm; left anterior neck adjacent to left parotid measuring 1.1 cm × 1.1 cm; and two other cervical lymph nodes measuring 1.2 cm × 0.9 cm and 1.0 cm × 0.6 cm, with no evidence of suppuration or abscess formation.

Table 1.

Test Results of the Patient’s Blood Sample for Her First Admission

Test	Result	Reference Range	Interpretation
White blood cell count	4.32 × 109/L	5.22–13.35 × 109/L	↓ Leukopenia
C-reactive protein (CRP)	11.2 mg/L	0–5 mg/L	↑ Elevated
Procalcitonin	0.43 µg/L	≤0.09 µg/L	↑ Mildly elevated
Erythrocyte sedimentation rate (ESR)	25 mm/hr	3–15 mm/hr	↑ Markedly elevated
Ferritin	29944.6 µg/L	13.7–78.8 µg/L	↑ Markedly elevated
Lactate dehydrogenase (LDH)	1140 U/L	192–321 U/L	↑ Elevated
Aspartate transaminase (AST)	63 U/L	21–44 U/L	↑ Elevated
Alanine transaminase (ALT)	28 U/L	9–25 U/L	Slightly ↑
Gamma-glutamyl transferase (GGT)	17 U/L	6–15 U/L	Slightly ↑

Notes: ↑: higher than reference range, ↓: lower than reference range.

Diagnostic lymph node biopsy showed histiocytic necrotizing lymphadenitis, in keeping with KFD (Figure 2A–C). In view of bicytopenia, high levels of lactate dehydrogenase (2615 U/L, normal 192–321 U/L) and ferritin (29944 µg/L, normal 13.7–78.8 µg/L) from a repeat blood test, a bone marrow aspiration was performed which showed hypocellular marrow with no malignant infiltrates. Skin biopsy showed interphase dermatitis (Figure 2D). Periodic Acid-Schiff (PAS) staining and Epstein-Barr encoding region (EBER) in situ hybridization on skin biopsy were negative.

Figure 2.

Hematoxylin and Eosin (H&E) staining of lymph node biopsy at (A) 40× and (B) 200× showed histiocytic necrotizing lymphadenitis with extensive lymph node necrosis with karyorrhectic debris rimmed by infiltrates consisting of immunoblasts, small lymphocytes, plasmacytoid dendritic cells and histiocytes. (C) CD123 staining with numerous plasmacytoid dendritic cells surrounding areas of necrosis at 100×. (D) Skin biopsy at 100× showed full thickness epidermal necrosis and adjacent basal vacuolar alteration with keratinocytes.

Treatment

The patient started on anti-inflammatory doses of non-steroidal anti-inflammatory drugs, 12.5 mg oral indomethacin twice a day (approximately 1 mg/kg), with omeprazole given for gastric protection. Her rashes were initially thought to be closely related to omeprazole, but it worsened despite cessation of omeprazole, making it less likely a drug-related rash. As fever did not respond to indomethacin, she was given anti-inflammatory doses of prednisolone (2 mg/kg). The fever subsided and there was improvement in the rashes within 24 hours of steroid initiation.

On outpatient review 1 week later, she was well with no further recurrence of fever or rashes with weaning doses of prednisolone (Figure 3). She received a total of four weeks of prednisolone. Her inflammatory markers also reflected improvement, with CRP at 2 mg/L, ferritin at 12841 µg/L and ESR at 9 mm/hr. The liver function test four months later returned normal results, with the alanine transaminase at 14 U/L (normal 9–25 U/L) and aspartate transaminase levels at 30 U/L (normal 21–44 U/L).

Figure 3.

Three months post-treatment, the lesions were largely resolved with only residual hyperpigmentation on her (A) face, (B) arms and (C) legs.

Recurrence

The patient was readmitted when she was 6 years old for 8 days of fever with no other localising symptoms. On examination, she had an isolated enlarged cervical lymph node on the right measuring about 1.5 cm × 1.5 cm. She had no other palpable lymphadenopathy, hepatosplenomegaly or joint swelling, and no rashes. Blood test results (Table 2) were similar to those from the first admission (Table 1). Infective markers were not significantly elevated, with mild leukopenia, CRP at 15.7 mg/L and procalcitonin 0.14 µg/L. Serology tests including urine and blood culture did not identify any bacterial or viral infection. Other infective screens including ASOT, Mycoplasma pneumoniae antibody, Bartonella serology, respiratory pathogens multiplex PCR, Cytomegalovirus IgM, EBV IgM, Dengue serology, Melioidosis PCR were negative. Anti-nuclear antibody and anti-double-stranded DNA antibodies were also repeated for this admission and were again negative. An ultrasound confirmed several cervical lymph nodes with the most prominent ones at the left submandibular intraglandular lymph nodes measuring 0.5 cm × 1.3 cm and 0.8 cm × 1.0 cm, the right anterior neck 0.9 cm × 1.9 cm and the left anterior neck 1.1 cm × 1.7 cm. There was no evidence of surrounding echogenic fat or lymph node suppuration.

Table 2.

Test Results of the Patient’s Blood Sample Upon Her Readmission

Test	Result	Reference Range	Interpretation
White blood cell count	2.34 × 109/L	5.22–13.35 × 109/L	↓ Leukopenia
C-reactive protein (CRP)	15.7 mg/L	0–5 mg/L	↑ Elevated
Procalcitonin	0.14 µg/L	≤0.09 µg/L	↑ Mildly elevated
Erythrocyte sedimentation rate (ESR)	58 mm/hr	3–15 mm/hr	↑ Markedly elevated
Ferritin	1976 µg/L	13.7–78.8 µg/L	↑ Markedly elevated
Lactate dehydrogenase (LDH)	831 U/L	192–321 U/L	↑ Elevated
Aspartate transaminase (AST)	66 U/L	21–44 U/L	↑ Elevated
Alanine transaminase (ALT)	26 U/L	9–25 U/L	Slightly ↑
Gamma-glutamyl transferase (GGT)	16 U/L	6–15 U/L	Slightly ↑

Notes: ↑: higher than reference range, ↓: lower than reference range.

The parents did not consent for a repeat diagnostic lymph node biopsy and KFD was diagnosed clinically. The patient was treated with prednisolone (1 mg/kg) with good response in the fever subsiding within 24 hours of initiation. On subsequent outpatient reviews, the prednisolone was tapered gradually with continued recovery. The patient was well in the latest review, 2 weeks after prednisolone was stopped.

Molecular Methods

Venous blood from the patient and the family was collected after written informed consent from her parents. High-resolution sequencing-based typing for 17 HLA loci was performed for the patient, her mother, maternal grandmother and unaffected father. Genomic DNA extracted from peripheral blood using Gentra Puregene Blood Kit (Qiagen, USA) was used to prepare the sequencing library using AlloSeq Tx17 (CareDx, USA). Sequencing was performed on the MiSeq system (Illumina, USA). Genotypes were assigned using AlloSeq Assign (CareDx, USA) with IPD-IMGT/HLA Database v3.47.

Next-generation sequencing was performed only on the patient’s DNA sample with TruSight One panel (Illumina Inc., San Diego, USA) on the MiSeq system. Reads were aligned to GRCh37 with Burrows-Wheeler Aligner, and variant calling was done with Genome Analysis Toolkit HaplotypeCaller. Variants were annotated using wANNOVAR and filtered for rare variants in the East and South Asian populations using Genome Aggregation Database (gnomAD) v2.1.1. Variants in genes with immune-related functions or reported in previous genetic studies on KFD patients were shortlisted for analysis. Missense variants were filtered further using in silico tools (Sorting Intolerant From Tolerant, MutationTaster and PolyPhen-2). The variants of interest were validated by Sanger sequencing, with inheritance status established by targeted Sanger sequencing of family samples. Variants were classified using the American College of Medical Genetics and Genomics and the Association for Molecular Pathology guidelines.¹⁸

Results

HLA Genotypes

Genotypes for 16 out of 17 HLA loci for the four family members were determined by AlloSeq. Except for HLA-DRB5 locus, the assigned genotypes for the 16 loci and the reported associations with other diseases are presented in Table 3. Fourteen alleles are shared by the three affected family members but are absent in the unaffected father. Five of them (C*08:02, DPA1*02:01, DQA1*02:01, DQB1*03:03 and DRB1*07:01) have reported associations with autoimmune inflammatory disorders.

Table 3.

HLA genotypes of the patient and family members

Gene	Patient	Grand-mother	Mother	Father	Disease association for shared allele
A	24:02:01:01	02:01:01:01	24:02:01:01	24:02:01:01	HIV protection19
	33:01:01:01	33:01:01:01	33:01:01:01	24:17:01:01	Severe dengue protection20
					Higher risk for terbinafine drug-induced liver injury21
B	14:02:01:01	14:02:01:01	14:02:01:01	15:02:01:01	Higher risk of aplastic anemia22
	18:01:01:17	18:01:01:52	52:01:01:01	18:01:01:17	HIV control23
					COVID-19 protection24
					Higher risk for trimethoprim-sulfamethoxazole drug-induced liver injury25
C	07:04:01:01	07:01:01:16	08:02:01:01	07:04:01:01	Higher risk of Graves’ orbitopathy26
	08:02:01:01	08:02:01:01	12:02:02:01	08:01:01:01	Chronic HIV infection27
					Milder COVID-1928
DPA1	02:01:01	02:01:01	01:03:01	01:03:01	Higher risk of systemic lupus erythematosus29
	04:01:01	04:01:01	02:01:01	04:01:01	Higher risk of diffuse cutaneous systemic sclerosis30
					Higher risk of COVID-19 hospitalization31
DPB1	14:01:01	14:01:01	14:01:01	04:01:01	Better response to Hepatitis B vaccine32
	296:01	296:01	04:01:01	296:01
DQA1	02:01:01	02:01:01	01:05:01	06:01:01	Higher risk of psoriasis33,34
	06:01:01	06:01:01	02:01:01	06:01:01	Lower risk of type 1 diabetes35, but higher risk when older36
					Higher risk of Parkinson’s disease37
DQB1	03:01:01	03:01:01	03:03:02	03:01:01	Higher risk of generalized pustular psoriasis38
	03:03:02	03:03:02	05:01:01	03:01:01	Lower risk of chronic hepatitis B/C39,40
					Higher risk of cutaneous malignant melanoma41
DRB1	07:01:01	07:01:01	07:01:01	12:02:01	Higher risk of psoriasis33,34
	12:02:01	12:02:01	10:01:01	12:02:01	Higher risk of bullous pemphigoid42
					Lower risk of rheumatoid arthritis43
					Lower risk of cytomegalovirus infection44
					Higher risk of immunodeficiency caused by HIV or tuberculosis45
					Higher risk for lapatinib-induced hepatotoxicity46
DRB3	03:01:03	03:01:03	not assigned	03:01:03
				03:01:03
DRB4	01:03:01N	01:03:01N	01:03:01N	not assigned
			01:03:01N
E	01:01:01	01:01:01	01:01:01	01:03:01:01	Higher risk of multiple sclerosis in individuals with infectious mononucleosis history47
	01:03:02:01	01:03:02:01	01:01:01	01:03:02:01	COVID-19 protection48
					Lower risk of hepatitis C in thalassemia patients and respond better to anti-HCV therapy49
F	01:01:01:09	01:01:01:09	01:01:01:09	01:01:01	01:01:02 associated with higher HLA-F mRNA expression50
	01:01:02	01:01:02	01:01:02	01:01:01
G	01:03:01	01:01:01	01:03:01:02	01:04:01	Lower risk of inflammatory bowel disease51
	01:04:01	01:03:01	01:04:01:01	01:04:01
H	02:12	01:01:01:01	02:12	not assigned	HLA-H is a pseudogene that is transcribed but not translated into a stable protein; it may regulate immune response via other HLA genes52
	02:12	02:12	02:12
MICA	011:01	008:04	009:01	018:01
	018:01	011:01	011:01	019:01
MICB	005:02	002:01	005:02	005:02
	014:01	005:02	005:02	014:01

Notes: Alleles shared by the three affected members but absent in the unaffected father are highlighted in red, and the reported disease associations with that shared allele are listed in the last column. Genotypes in bold have mismatches in the exon and were determined based on the closest match. N = null allele

Genes Related to Immunity

Targeted sequencing of the patient’s DNA with the TruSight One panel, which includes 4318 Mendelian genes, identified five damaging variants with frequencies of less than 1% in East and South Asian populations in genes associated with immune functions (Table 4). The presence of the variants in her family members was investigated by Sanger sequencing. Only three variants (FGFR3, NRP2 and PCM1) correlated with the KFD status in this family. All are missense variants predicted to be deleterious or damaging by in silico tools.

Table 4.

Sanger Sequencing Results for Rare Variants Identified from TruSight One

Gene Symbol	Variant Nomenclature	Variant Frequency (East/South Asian)	Gene’s Immune Function	Variant Type	Validated in
					Patient	Mother	Grandmother	Father
CCN2	c.1037delG p.Gly346Glufs*9 (NM_001901.4)	0.04%/0.01%	Regulates inflammation	Frameshift	−/+	−/+	−/−	−/−
FGFR3	c.967C>T p.Leu323Phe (NM_001163213.2)	Not available	Regulates inflammation	Missense	−/+	−/+	−/+	−/−
HARS1	c.506G>A p.Arg169Gln (NM_002109.6)	0%/0%	Regulates immune cells migration	Missense	−/+	−/−	−/−	−/+
NRP2	c.470A>T p.Asn157Ile (NM_003872.3)	0.02%/0%	Regulates immune cells migration	Missense	−/+	−/+	−/+	−/−
PCM1	c.4538C>G p.Thr1513Ser (NM_006197.4)	0.07%/0%	Regulates autophagy and pyroptosis	Missense	−/+	−/+	−/+	−/−

Notes: Individuals who carry the variant were highlighted in grey. Variant in alleles were represented by +: present and −: absent.

Genetic variants in C1a, C1q, C4 and MEFV were previously reported in KFD patients. We found two heterozygous variants in C1QC and C5AR2 for genes of the classical complement pathways. However, they did not correlate with disease symptoms within the family (Table 5). Instead, the grandmother was found to have a homozygous frameshift deletion in C1QC. This 17-nucleotide deletion in exon 2 results in a premature stop codon 69 amino acids downstream, which is likely to cause C1q deficiency and affect the complement pathway. Regardless, our patient does not carry this deletion while her mother is heterozygous. We also found a heterozygous MEFV variant in our patient. This single nucleotide variant is inherited from the mother but not present in the grandmother and father (Table 5). Although some of these variants could be pathogenic, none of them segregates with the disease phenotype.

Table 5.

Sanger Sequencing Results for Variants Reported to Be Associated with KFD

Gene Symbol	Variant Nomenclature	Variant Frequency (East/South Asian)	Variant Type	Validated in
				Patient	Mother	Grandmother	Father
C5AR2	c.899G>A p.Arg300His (NM_001271749.2)	3.23%/0.01%	Missense	−/+	−/−	−/−	−/+
C1QC	c.559G>A p.Val187Ile (NM_172369.5)	2.33%/0.11%	Missense	−/+	−/+	−/−	−/−
C1QC	c.106_122del p.Pro36Alafs*69 (NM_172369.5)	Not available	Frameshift	−/−	−/+	+/+	−/−
MEFV	c.910G>A p.Gly304Arg (NM_000243.3)	1.81%/0.11%	Missense/splicing	−/+	−/+	−/−	−/−

Notes: Individuals who carry the variant were highlighted in grey. Variant in alleles were represented by +: present and −: absent.

Discussion

Other HLA Studies in Patients with KFD

There were four previous studies that investigated the HLA genotypes of sibling pairs with KFD. All found at least one shared allele between the sibling pair, yet none of them were common across all four sibling pairs (Table 6). An earlier Japanese study reported a 15–20% higher frequency of DPA1*01 and DPB1*02:02 alleles in patients compared to controls.¹² These two alleles were not present in the three affected individuals of our study, but DPB1*02:02 was found in a Japanese sibling pair from another study.¹¹ Due to the use of traditional HLA typing methods in some older studies with only one field resolution, inter-studies comparison could only be done at that resolution. The alleles A*02, B*35, C*07 and DRB*08 were found in two or more sibling pairs but not in the three affected family members of our study. A KFD parent-child pair has also been reported but their HLA genotypes were not investigated.⁷

Table 6.

HLA Studies on Siblings in Familial Cases

HLA	Isoda et al (2023)11		Quadir et al (2020)10	Stasiuk et al (2011)9		Amir et al (2002)8
	Brother	Sister	Identical Twin Sisters	Sister 1	Sister 2	Sisters
A	02:07:01 24:02:01	02:07:01 24:02:02	02:01 03:02	02	02 31	31
B	40:02:01 46:01:01	40:02:01 46:01:02	08:01 35:08	35 39	39 51	35 49
C	01:02:01 03:04:01	01:02:01 03:04:02	04:01 07:02	Cw4 Cw7	Cw7 Cw15	04 07
DPA1	01:03:01 02:02:02	02:02:02
DPB1	02:02:01 03:01:01	02:02:01 05:01:01
DQA1	01:03:01 03:03:01	01:03:01 03:01:01	02:01 05:01
DQB1	04:02:01 06:01:01	03:02:01 06:01:01	02:01 02:02	03	03 04:02
DRB1	04:10:03 08:03:02	08:02:01 08:03:02	03:01 07:01	08 14	08 14
DRB3			02:02
DRB4	01:03:01		01:03

Notes: For comparison of HLA genotypes from siblings, some alleles were not reported or only had one field resolution (2002 and 2011 publications).

Among the five HLA loci (C*08:02, DPA1*02:01, DQA1*02:01, DQB1*03:03 and DRB1*07:01) that are shared by affected individuals and absent in the unaffected individuals of our patient’s family, DQA1*02:01 and DRB1*07:01 were also found in a pair of affected twins from Australia.¹⁰ Both alleles have been associated with psoriasis and systemic sclerosis which are inflammatory diseases with both genetic and environmental factors, similar to KFD.^53,54 Incidentally, these two alleles were also associated with drug reactions such as asparaginase hypersensitivity in paediatric acute lymphoblastic leukemia patients and lapatinib-induced hepatotoxicity in breast cancer patients.^46,55 These associations suggest an overactive immune reaction, which is also observed in KFD.

Immune-Related Genes Associated with KFD

Previous studies have found variants linked to C1s, C1q and C4 deficiencies in KFD patients.^13,14,16 These genes are involved in the classical complement pathway of the innate immunity that could cause inflammation, suggesting that mutations in other inflammatory immune genes could also contribute to KFD. Our study identified three variants in genes of the complement pathway but none of them segregates with KFD for the family (Table 5). For the two missense variants, the one in C5AR2 was inherited from the unaffected father while the C1QC variant was classified as “benign” on ClinVar by two submissions with no phenotype listed. The last variant is a novel 17-nucleotide deletion in C1QC (Pro36Alafs*69) which can affect C1qc expression. It was found in two of the three affected individuals of the family and absent in the unaffected father. Although it is absent in the young patient, the variant could be a contributing factor to the disease in the mother and grandmother as suggested by previous findings.

Variants in MEFV have also been reported in patients diagnosed with KFD.^15,17 This gene encodes pyrin which regulates inflammasomes activation, resulting in the release of interleukin-18 (IL-18). A KFD patient carrying compound heterozygous variants (Pro369Ser and Arg408Gln) was successfully treated with Colchicine, by inhibiting the inflammasomes and reducing the level of IL-18.¹⁵ The variant we identified in the patient (Gly304Arg) has not been reported in KFD but was found in cases of familial Mediterranean fever.⁵⁶ It is located at the end of exon 2, which can cause aberrant splicing by enhancing exon 2 skipping, resulting in decreased expression of the normal mRNA. There are 14 submissions of this variant in ClinVar with conflicting classifications of pathogenicity. Six of them were reported with conditions, either Familial Mediterranean fever, MEFV-related disorder or auto-inflammatory syndrome. However, the pathogenicity classifications include a mixture of benign, likely benign and uncertain significance. Since the variant is present in two affected individuals and reported to affect splicing, it is also a possible factor contributing to KFD.

In addition to these previously reported gene associations, we have identified three other potential variants in immune-related genes which are predicted to be damaging by in silico tools. They also correlate with the KFD status in the family (Table 4).

The only variant previously unreported is found in FGFR3 (NM_001163213.2, c.967C>T, p.Leu323Phe). This variant is located in exon 8 which is not present in the MANE Select transcript and most other isoforms. However, the transcript (NM_001163213.2) with exon 8 has the highest TRIFID functional isoform scores assigned by APPRIS and the only one annotated as a principal isoform among the rest, suggesting its importance. FGFR3 is a receptor for fibroblast growth factors and is known to cause skeletal developmental disorders when mutated. The receptors are also associated with cancer and autoimmune disorders due to their involvement in inflammatory regulation.

The missense variant in exon 4 of NRP2 (NM_003872.3, c.470A>T, p.Asn157Ile) has a frequency of 0.001% and is only found in East Asians at 0.016%. This variant has two submissions in ClinVar with, uncertain significance to NRP2-related disorder. NRP2 encodes neuropilin, a non-tyrosine kinase glycoprotein expressed on the surface of immune and other cells. It can regulate physiological processes as well as immune responses.

Lastly, the missense variant in exon 28 of PCM1 (NM_006197.4, c.4538C>G, p.Thr1513Ser) has a frequency of 0.005% across all populations and 0.07% in East Asians. It is also present at 0.2% in the Singapore population according to SG10K database, with a higher frequency in Malays compared to other ethnic groups.⁵⁷ PCM1 is a component of centriolar satellites but has shown involvement in immune responses, such as regulating autophagy and pyroptosis.^58,59

As KFD is characterized by hyperactive immune response, immune-regulatory genes are likely to be involved in the disease’s manifestation. While we identified two variants in KFD-associated genes and three variants in genes not previously associated with KFD, their specific impact on immune dysregulation and KFD pathogenesis remains to be investigated with functional studies.

Other Genetic Studies on KFD

Genome sequencing studies by other groups have unveiled variants in other genes from patients diagnosed with KFD. Fourteen genes were selected by a Thai group from whole-exome sequencing of four KFD patients.⁶⁰ The genes were associated with various cancers, olfaction, and osteoblast differentiation. Among those, OR8U8 which encodes an olfactory receptor was highlighted by another study with a patient who also carried eight polymorphisms in the gene.⁶¹ Unfortunately, these 14 genes were not in the TSO panel used for our patient.

Apart from DNA sequencing, RNA sequencing has also provided some insights into the disease. A transcriptomic study comparing 15 KFD and 9 non-KFD lymphadenitis cases identified the type I interferon response pathway as a mediator of the disease.⁶² Perforin-2, which is a regulator of type I interferon response, was found to be overexpressed in KFD.⁶³ Another transcriptomic study reported altered gene expression for genes involving the immune system, chromatin remodelling, and transcription regulation.⁶⁰ These processes could influence the immune reaction, resulting in the overactive immune response seen in KFD. Transcriptomic data can provide information complementary to genomic data in identifying potential contributing factors to KFD.

Conclusion

Using current next-generation sequencing technology that can determine HLA genotypes at a high resolution of up to four fields, we found five HLA alleles and also three variants in immune-related genes that could be associated with the disease based on published data and in silico pathogenicity prediction. More evidence from other familial cases or functional studies is needed to support our findings.

The rarity of KFD cases and limited resources of some institutions make it challenging for meaningful analysis and cross-validation. Hence, global data sharing is important for integrative data analysis, which also minimizes ethnic bias. Families with both affected and unaffected members will be especially useful for the validation of results.

Funding Statement

This work was supported by the National Research Foundation Singapore under its NMRC Centre Grant Program (NMRC Project No. NMRC/CG1/006/2021) administered by the Singapore Ministry of Health’s National Medical Research Council.

Disclosure

The authors report no conflicts of interest in this work.