Favorable preliminary experience with etanercept in two patients with the hyperimmunoglobulinemia D and periodic fever syndrome

Authors


Abstract

Objective

The hyperimmunoglobulinemia D and periodic fever syndrome (HIDS; MIM 260920) is caused by recessive mutations in the mevalonate kinase gene (MVK), which encodes an enzyme involved in cholesterol and nonsterol isoprenoid biosynthesis. HIDS is characterized by persistently elevated polyclonal IgD and recurrent febrile episodes. Although abnormalities in tumor necrosis factor α (TNFα) are not the primary cause of HIDS, plasma TNFα levels are elevated in HIDS patients during attacks and thus may be a therapeutic target. This study assessed the effects of etanercept, a soluble p75 TNFα receptor-Fc fusion protein, in 2 patients with HIDS.

Methods

We performed biochemical and molecular genetic analyses on 2 girls with periodic episodes of fever, skin rash, abdominal pain, and arthralgia, of whom 1 had elevated levels of serum IgD. After the diagnosis of HIDS was made, treatment with etanercept was initiated in both patients. Clinical response was recorded in a standardized diary, and serum levels of cytokines and their decoy receptors were serially measured in 1 of the 2 patients.

Results

Urinary mevalonate levels were elevated in both girls. Patient 1 was heterozygous for a known MVK missense mutation (V377I) and a novel mutation that led to skipping of exon 3. Patient 2 was found to have V377I and a new missense mutation, S329R. Neither patient had mutations in TNFRSF1A or MEFV, the genes for the TNF receptor–associated periodic syndrome and familial Mediterranean fever, respectively. Etanercept reduced the frequency and severity of symptoms in both patients, whereas the levels of serum IgD and urine mevalonate remained unchanged.

Conclusion

Our favorable experience with etanercept for the treatment of HIDS suggests that further investigation of this therapy is warranted.

The hyperimmunoglobulinemia D and periodic fever syndrome (HIDS; MIM 260920) is a rare autosomal recessive disease that is characterized by constantly elevated polyclonal IgD in most patients and recurrent febrile attacks. The attacks are frequently accompanied by chills, headache, bilateral cervical lymphadenopathy, arthritis, abdominal pain, and diarrhea. The median age at onset of HIDS is 0.5 years, and the frequent and regular recurrence of attacks over many years is often debilitating (1).

Recently, 2 independent groups identified the gene responsible for HIDS as MVK, which encodes mevalonate kinase (2, 3). Mevalonate kinase catalyzes a key early step in a pathway that produces cholesterol and a host of other bioactive molecules. A near-complete deficiency of this enzyme had been previously shown to cause mevalonic aciduria, which shares some characteristics with HIDS but has a more severe clinical presentation and results in failure to thrive, mental retardation, hypotonia, and dysmorphic facies (4).

HIDS and mevalonic aciduria are caused by variant MVK alleles, and nearly all patients with HIDS have at least 1 copy of the V377I mutation of MVK (5, 6). Despite the identification of the responsible gene, the precise metabolic and molecular pathophysiology of HIDS still remains unclear. Various therapeutic approaches have been undertaken in an attempt to suppress inflammatory episodes and to modify altered metabolic systems, but there is no effective treatment yet for HIDS (1).

Some features of HIDS, such as fever and an acute-phase response, are mediated by proinflammatory cytokines such as interleukin-6 (IL-6), tumor necrosis factor α (TNFα), and IL-1. Drenth et al showed significant elevations in the circulating levels of IL-6 and, albeit very modest, in TNFα, together with elevations of the natural cytokine inhibitors IL-1 receptor antagonist (IL-1Ra), soluble TNF receptor superfamily 1A (sTNFRSF1A), and sTNFRSF1B during febrile attacks in 22 HIDS patients (7). They also showed an increased lipopolysaccharide-stimulated ex vivo production of IL-1β and TNFα in whole blood during febrile attacks. We therefore speculated that the dysregulated production of these proinflammatory cytokines contributes to the clinical manifestations of HIDS, and that, given its pleiotropic effects, TNFα plays a central role.

Herein we describe 2 patients with HIDS who were diagnosed on the basis of clinical, immunologic, metabolic, and molecular findings. Therapy with etanercept (Enbrel; Immunex, Seattle, WA), a fusion protein consisting of 2 chains of the recombinant human TNFα receptor p75 monomer fused with the Fc domain of human IgG1 (TNFR-Fc fusion protein), led to significant reduction in the frequency and severity of febrile attacks in both patients.

PATIENTS AND METHODS

Analysis of mutations.

Mutational analyses of MEFV and TNFRSF1A were conducted as previously described (8, 9). MVK complementary DNA (cDNA) sequencing was carried out using published primer sets (3). Exons are denoted by the numbering system of Houten et al (5). The V377I mutation was confirmed by digestion of genomic DNA with Bsm AI. We also designed an Sph I restriction endonuclease assay for the S329R substitution, using a primer pair that creates a cleavage site in the wild-type amplicon.

The MVK exon 3 deletion was initially detected by amplification of cDNA with the primer set 29/535 as described by Drenth et al (3), and then further confirmed with nested primers. The amplicons were ligated into the pCR2.1 vector using the TOPO TA Cloning Kit (Invitrogen, San Diego, CA), and were sequenced with dye-primer chemistry (Amersham, Arlington Heights, IL). MVK exon 3 splice sites were also sequenced. All primer sequences are available on request.

Mevalonate and cytokine assays.

Ethylacetate extraction of mevalonate (as the lactone) from urine was performed using [2H3]-mevalonate as an isotopic internal standard and N-methyl-N-(tert-butyldimethylsilyl)trifluoracetamide + 1% tert-butyldimethylchlorosilane as a derivatizing agent. Gas chromatography was performed on a 0.20 mm × 25 meter methylsilicone (0.33 μm phase) column with quantification by selected-ion mass spectrometry at 317/321 m/z in the electron impact mode, at 70 eV and a source temperature of 200°C. Values from 6 healthy subjects (age range 0–15 years) were used to obtain a normal range for urinary mevalonate levels (mean ± SD 0.40 ± 0.18 mg/gm creatinine).

From patient 1, serum specimens were collected at 5 different time points and were stored at −80°C. Measurements of cytokine levels in the sera were performed by routine enzyme-linked immunosorbent assay using the Clinical Services Program (National Cancer Institute, Frederick, MD).

Clinical evaluation.

The patients' mothers were asked to keep a daily diary to register the presence or absence of fever and other signs or symptoms. HIDS attacks were defined as fever (≥38.5°C, not caused by infection) together with one or a combination of the following symptoms: abdominal distress (pain, vomiting, diarrhea), joint involvement (arthralgia, arthritis), skin lesions, and/or lymphadenopathy. Usually, fever was present in conjunction with these symptoms.

RESULTS

Clinical presentation, molecular genetic analysis, and clinical response to etanercept.

Patient 1.

Patient 1 is the second daughter of parents of Irish and Albanian origin. At 6 weeks of age, she developed fever, erythematous maculopapular rash, and diarrhea, which lasted for 7 days, followed by spontaneous resolution. Since then, she has experienced recurrent and self-limited febrile attacks accompanied by erythematous maculopapular rash (Figure 1A), abdominal pain and diarrhea, generalized myalgia and arthralgia, tender and swollen cervical lymph nodes, and occasional oral ulcers and headache, with significant elevation in the acute-phase reactants. Attacks occurred every 2–6 weeks, lasting for 4–10 days, and were always followed by spontaneous and complete resolution of symptoms. She has had normal physical and intellectual development for her age. No one in the family had a similar illness.

Figure 1.

Clinical findings in a patient with hyperimmunoglobulinemia D and periodic fever syndrome (HIDS). A, Erythematosus macular and papular rash distributed over the lower and upper extremities and trunk was observed in patient 1 during a HIDS attack. The photograph was made by the parents and was reproduced with permission. B, Daily measurements of the urinary mevalonate levels in patient 1 revealed elevated baseline levels of urinary mevalonate and a significant spike during a typical febrile attack (normal range mean ± SD 0.40 ± 0.18 mg/gm creatinine).

In December 1998 at the age of 10 years, she was referred to the Warren Grant Magnuson Clinical Center of the NIH, where she was noted to have persistent elevations in the levels of serum IgD (88.5 mg/dl and 54.4 mg/dl, normal range 0.0–14.0 mg/dl; Mayo Medical Labs, Rochester, MN). Furthermore, her urinary levels of mevalonate, a substrate of mevalonate kinase, were found to be elevated modestly at baseline and elevated significantly during febrile attacks (Figure 1B).

Molecular genetic analysis demonstrated that patient 1 was a compound heterozygote for a known missense mutation (V377I) and for a mutation leading to skipping of exon 3 in the cDNA (Figure 2A), inducing a frameshift in exon 4 and, consequently, leading to a downstream stop codon. Mutational screenings for MEFV and TNFRSF1A were negative.

Figure 2.

Molecular genetic findings in the study patients. A, Agarose gel electrophoresis of polymerase chain reaction (PCR) products of MVK cDNA, obtained using oligonucleotides 29 and 535 as described by Drenth et al (3), shows the normal 258-bp alleles in samples from healthy controls (C), as well as both the normal 258-bp allele and the mutant 110-bp allele in the sample from patient 1 (P). Subsequent cDNA sequencing confirmed that this 148-bp deletion in cDNA corresponds to exon 3 (according to the exon numbering of Houten et al [5]). Sequence analysis at the genomic level over 50 bp upstream or downstream of MVK exon 3, however, did not identify a responsible mutation that leads to exon skipping. The molecular size marker (L) was a 100-bp ladder. The ddH2O line represents PCR reaction with double-distilled water added, but not with DNA, indicating that there was no background amplification. B, In patient 2, direct genomic sequencing of MVK demonstrates the presence of a S329R missense mutation. Heterozygosity for the missense mutation occurs at nucleotide 987 (C→A), changing the serine at codon 329 to an arginine. C, Homozygosity for the normal allele. In B and C, the relevant nucleotide is indicated by the arrow.

After informed consent was obtained, etanercept was started at the dose recommended for the treatment of juvenile chronic arthritis (0.4 mg/kg subcutaneously, twice weekly), in August 1999. During the period from August 1999 to February 2000, the number of days of HIDS attacks was about the same as was estimated for the pretreatment period (30 days/6 months). During this period, there was some difficulty in achieving the optimal balance between the benefit of treatment and its side effects, which were mainly upper respiratory infections, and the dosage frequency varied between once per week and 3 times per week. However, as shown in Figure 3A, beginning in March 2000, there was a marked reduction in the number of attack days. Furthermore, the severity of each attack was noted to be decreased. She currently takes etanercept once weekly. In August 2001, the etanercept injections were omitted for 5 weeks and the patient subsequently experienced a full-blown attack.

Figure 3.

Clinical effects of etanercept. Changing trends are evident in the numbers of days sick with hyperimmunoglobulinemia D and periodic fever syndrome (HIDS) attacks in relation to etanercept use in each 6-month period in patient 1 (A) and patient 2 (B). HIDS attacks were defined as days during which the patients had fever (≥38.5°C, not caused by infection) together with one or a combination of the following symptoms: abdominal distress (pain, vomiting, diarrhea), joint involvement (arthralgia, arthritis), skin lesions, and/or lymphadenopathy. Such clinical information was obtained from daily diaries completed by the patients' mothers. = During this period, the patient did not take etanercept for 5 weeks due to noncompliance, and this was followed by a full-blown attack. No statistical comparison was made to compare the number of days sick with HIDS before and during etanercept treatment.

Patient 2.

Patient 2 is the first daughter of parents of western European (Italian, French, German, and Spanish) origin, who, since 3 days of life, has experienced recurrent fevers accompanied by leg pain, erythematous maculopapular eruption, abdominal bloating/pain and diarrhea, headache, throat pain, and lymphadenopathy, occurring every 7–9 days and lasting 48–72 hours, followed by spontaneous resolution. When she was referred to the NIH clinic in September 1999, her serum IgD level was within the normal range, but her urinary mevalonate levels were found to be significantly elevated.

Molecular genetic analysis of patient 2 revealed compound heterozygosity for the V377I mutation and a heretofore-unknown missense mutation, S329R (Figure 2B). S329R was not detected in any of 120 chromosomes from white control subjects. No mutations were found in MEFV or TNFRSF1A.

After informed consent was obtained, twice weekly dosing of etanercept was started in August 1999, at 0.4 mg/kg subcutaneously. Similar to patient 1, the number of HIDS attacks (Figure 3B) and the severity of HIDS attacks significantly decreased from that observed prior to the initiation of etanercept. In contrast to patient 1, the optimal balance between the benefit and side effects of etanercept was achieved relatively quickly. The serum IgD levels in patient 2 remain within the normal range.

Effect of etanercept on the systemic cytokine milieu.

We measured the levels of cytokines and soluble cytokine receptors (IL-1β, IL-6, TNFα, IL-1Ra, sTNFRSF1A, and sTNFRSF1B) in the serum samples obtained from patient 1 at 5 different time points throughout the course of treatment (Figure 4). All 5 samples were obtained when she was asymptomatic. Gradual normalization of the C-reactive protein (CRP) level was observed after etanercept was started. Although the levels of IL-1Ra, sTNFRSF1A, IL-1β (data not shown), and IL-6 (data not shown) were within the normal range, the levels of TNFα and sTNFRSF1B increased after etanercept was started. The increases in immunoreactive TNFα and sTNFRSF1B in patient 1 are likely to represent the formation of complexes of TNFα with etanercept, and thus document compliance with the regimen.

Figure 4.

Changes in the erythrocyte sedimentation rate (ESR), levels of C-reactive protein (CRP), interleukin-1 receptor antagonist (IL-1ra), tumor necrosis factor α (TNFα), soluble TNF receptor superfamily 1A (sTNFRSF1A), and sTNFRSF1B in the serum before and during etanercept therapy in patient 1. Serum specimens were collected at 5 different time points. The first sample (□) was obtained before the initiation of etanercept. The second (░), third (░), and fourth (░) samples were obtained in between attacks at 1, 3, and 12 months, respectively, after etanercept was started, while dosing was on a twice-weekly schedule. The fifth sample (▪) was obtained at 20 months after etanercept was started, while the dosing was on a once-weekly schedule. Within 24 hours after the fifth sample was drawn, patient 1 developed a few mild and short-lasting constitutional symptoms. Normal ranges for these values are as follows: IL-1ra 106–1,552 pg/ml, TNFα <15.6 pg/ml, sTNFRSF1A 749–1,966 pg/ml, sTNFRSF1B 1,003–3,170 pg/ml, and ESR 0–42 mm/hour. Levels of sTNFRSF1B in the second to the fifth samples were out of scale (>5,000 pg/ml) and were therefore cut at 5,000 pg/ml for the graph. This increase in immunoreactivity of sTNFRSF1B in patient 1 likely represents the recognition of not only sTNFRSF1B, but also the sTNFRSF1B moiety of etanercept by this assay, and thus documents the treatment compliance of patient 1 at the time that these blood samples were obtained. Levels of IL-1β and IL-6 were within normal limits (<15.6 pg/ml) in all samples and are therefore not shown. † = Level of CRP in the fourth sample was below the lowest detectable level (<0.4 mg/dl). ‡ = Level of CRP in the fifth sample was 13.5 mg/dl.

Within 24 hours after the fifth blood sample was obtained, patient 1 developed mild fatigue and slight malaise without fever. Although this was not recorded as a HIDS attack because of the absence of fever or other characteristic manifestations of HIDS attacks, patient 1 indicated that these symptoms were identical to the prodromal symptoms that she usually experiences before developing full-blown HIDS attacks, and, consistent with this assessment, significant elevation of the erythrocyte sedimentation rate and CRP level, as well as elevation of serum IL-1Ra (although still within the normal range) was observed. However, neither serum IL-6 nor IL-1β was elevated (data not shown), and the sTNFRSF1A levels, which are not affected by exogenous TNFR-Fc, were mildly elevated but within the normal range.

DISCUSSION

In both patients described in this report, etanercept therapy significantly reduced the frequency of HIDS episodes and ameliorated clinical manifestations during the attacks. The clinical effectiveness of etanercept has been previously demonstrated by the partial response observed in a HIDS patient who had a TNFRSF1A mutation (10). Neither patient in the present report had mutations in either TNFRSF1A or MEFV. These findings suggest that neutralization of TNFα may represent a novel and effective treatment for HIDS.

We also identified 2 new disease-associated alleles in the MVK gene. Patient 1 was found to be a compound heterozygote for a missense mutation, V377I, and an unidentified mutation leading to skipping of exon 3 in the cDNA. Patient 2 was a compound heterozygote for the missense mutations V377I and S329R. V377I is the most frequent mutation in HIDS (5, 6), but patient 2 is the first to be reported with the S329R allele. Exon skipping has been observed in only 2 patients, for exon 2 and exon 5 (3, 6). Houten et al recently showed the presence of 2 rather frequent natural splice variants (skipping of exon 4 or exons 4 and 5) and 1 rare natural splice variant (skipping of exon 8) of MVK (5), but skipping of exon 3, which we found in patient 1, has not been observed previously either in HIDS patients or in the 120 white control chromosomes that we have previously studied. We could not identify a mutation responsible for this altered splicing through the sequence analysis at the genomic level in the 50 bp upstream or downstream of MVK exon 3 splice sites.

Levels of cytokines and soluble cytokine receptors in serum were measured in patient 1 in an attempt to evaluate the effects of etanercept on the systemic cytokine milieu and to relate this to the clinical response. Unexpectedly, the fifth blood sample, which was, coincidentally, drawn shortly before patient 1 developed mild and short-lasting constitutional symptoms, revealed significant elevation of acute-phase reactants, IL-1Ra, and, to a lesser degree, sTNFRSF1A. This suggests that similar subclinical and afebrile “attacks” may continue to occur despite treatment with etanercept. This also underscores yet-unknown mechanisms through which mutations in MVK cause periodic inflammation.

The mechanism by which mutations in MVK cause febrile attacks is not clear. Mevalonate kinase, a product of MVK, catalyzes the conversion of mevalonate to 5-phosphomevalonic acid in the biosynthesis of cholesterol and nonsterol isoprenoid compounds. Decreased mevalonate kinase activity, which fever aggravates (11), leads to accumulation of its substrate, mevalonate. Some clinical features in patients with mevalonic aciduria may be directly due to the accumulation of mevalonate. However, in one report, 2 patients with mevalonic aciduria developed severe febrile attacks following treatment with lovastatin, despite an initial decrease of mevalonate levels in the serum and urine (4), and it is therefore less likely that increased production of mevalonate itself is directly involved in the pathogenesis of febrile attacks.

Decreased mevalonate kinase activity also leads to decreased production of molecules involved in prenylation, which is the posttranslational modification of proteins with isoprenoids such as farnesyl and/or geranyl moieties. Particularly interesting cellular proteins that undergo prenylation are those from the Ras, Rho/Rac, and Rab families of small GTP binding proteins (12), which are important intermediates in signal transduction. Impaired prenylation, particularly of those proteins providing regulatory control over ligand-induced cellular activation, may lower the threshold for proinflammatory cytokine production. A recent study documents increased IL-1 production by HIDS leukocytes in vitro (13), but a direct link between disrupted prenylation and cytokine production or febrile attacks is yet to be elucidated.

In summary, we have described 2 HIDS patients in whom we identified 2 new disease-associated alleles, and in whom etanercept successfully attenuated the clinical manifestations of febrile attacks. Our findings therefore further support the important role that TNFα plays in HIDS, and indicate that therapies targeting TNFα have the potential to be a cornerstone in the treatment of HIDS. Our findings also underscore the importance of careful monitoring and dose adjustment to achieve an optimal benefit–risk balance with such therapies.

Acknowledgements

The invaluable contribution of the nurses of the 9 West Day Hospital, Warren Magnuson Clinical Center, is acknowledged. We would also like to express our gratitude to the study patients and their referring physicians.

Ancillary