Discordance between disease activity and long-term outcomes in DNASE1L3 deficiency: a multicentre longitudinal cohort study

Abstract

Objectives

DNASE1L3 deficiency is a rare monogenic cause of early-onset SLE and hypocomplementaemic urticarial vasculitis syndrome (HUVS). While early disease manifestations are increasingly recognised, longitudinal data extending into adulthood remain scarce. In this study, we described the long-term disease trajectory, damage accrual and mortality across the largest DNASE1L3 deficiency cohort reported to date.

Methods

We conducted a multicentre retrospective longitudinal cohort study of patients with genetically confirmed DNASE1L3 deficiency followed at paediatric rheumatology tertiary centres in Oman and the UAE. Primary outcomes included damage accrual, flare rates, hospitalisation frequency, serious infections and mortality. Longitudinal comparisons were made between the last paediatric assessment and most recent adult follow-up.

Results

57 patients from 22 families were enrolled, most carrying the homozygous DNASE1L3 c.643delT loss of function variant. Median age at diagnosis was 5 (1.75–13) years and median disease duration was 9 (1–22) years. Initial phenotypes comprised of SLE in 36 patients (63%) and HUVS in 21 (37%). Despite a significant reduction in disease activity (SLE Disease Activity Index scores) over the follow-up period (9 vs 4; p<0.001), this was not accompanied by clear improvement in long-term outcomes. Mean Systemic Lupus Interntational Collaborating Clinics (SLICC)/American College of Rheumatology Damage Index increased numerically from 0.37±0.85 to 0.72±1.17; p=0.295. The proportion with lupus nephritis increased from 34% to 46% (p=0.07) and pulmonary involvement from 3.7% versus 11% at follow-up (p=0.289), osteoporosis increased from 23% to 41% (p=0.065) and mortality was numerically higher following transition, increasing from 8.8% to 13.5% (p=0.125).

Conclusions

In the largest DNASE1L3 deficiency cohort, reduction in clinical disease activity does not prevent progressive organ damage. These findings suggest that apparent disease quiescence may mask ongoing subclinical injury, underscoring the need for sustained long-term surveillance and optimised transitional care strategies. However, the findings should be interpreted with caution in light of the small sample size in this ultra-rare disease.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• DNASE1L3 deficiency is a rare, fully penetrant autosomal-recessive cause of monogenic SLE and hypocomplementaemic urticarial vasculitis syndrome, typically presenting in early childhood with treatment-refractory disease. While previous studies have defined the paediatric phenotype and its severity relative to sporadic childhood-onset SLE, longitudinal outcomes after transition to adult care remain poorly characterised.

WHAT THIS STUDY ADDS

• This study provides the largest longitudinal assessment of genetically confirmed DNASE1L3 deficiency to date and identifies a clinically important discordance between improved disease activity scores and ongoing adult morbidity. Despite lower SLE Disease Activity Index (SLEDAI) scores over time, patients showed increasing damage burden, greater renal and skeletal involvement, emerging cardiopulmonary complications and continued mortality after transition to adult care.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings challenge the assumption that reductions in SLEDAI reflect adequately controlled disease in monogenic lupus and suggest that current disease activity metrics may inadequately capture subclinical inflammatory burden in DNASE1L3 deficiency. They support continued adult surveillance after transition from paediatric care. Future therapeutic strategies, including emerging enzyme-replacement biologics, may need to target prevention of end-organ damage rather than symptom-based disease activity alone.

Introduction

SLE is a clinically heterogeneous autoimmune disease driven by defective handling of self-nucleic acids and consequent loss of immune tolerance. While most cases arise through complex polygenic mechanisms, rare monogenic forms of lupus have provided important insights into disease pathogenesis. More than 30 genes have now been implicated in monogenic lupus. Among these, pathogenic variants in DNASE1L3 have emerged as a clinically distinct form of monogenic SLE.¹

DNASE1L3 encodes a secreted endonuclease that degrades extracellular chromatin derived from apoptotic cells, microparticles and neutrophil extracellular traps. Homozygous loss-of-function variants impair clearance of extracellular DNA, promoting persistent immunogenic chromatin exposure, anti-double-stranded (dsDNA) and anti-chromatin autoantibody formation, immune complex-mediated inflammation and a lupus-like disease.² Biallelic DNASE1L3 mutations cause autosomal-recessive monogenic SLE, and the same variants have subsequently been identified in patients with hypocomplementaemic urticarial vasculitis syndrome (HUVS), establishing that DNASE1L3 deficiency encompasses a phenotypic spectrum ranging from HUVS without fulfilment of SLE classification criteria to fully expressed SLE.^{3 4} The hallmark clinical features include very early-onset disease, persistent hypocomplementaemia, high-titre anti-dsDNA antibodies, urticarial vasculitis, lupus nephritis and, in some cases, pulmonary haemorrhage.⁵

Despite growing recognition of DNASE1L3 deficiency as a cause of monogenic lupus, its long-term clinical course remains incompletely described. Existing reports have mainly focused on early disease manifestations in small cohorts, with limited data on outcomes beyond childhood.^{3 5 6} We previously described the largest genetically confirmed DNASE1L3 deficiency cohort reported at that time.⁵ This expanded study examines long-term disease trajectory, damage accrual and mortality after transition to adult care.

Methods

We conducted a multicentre, retrospective, longitudinal cohort study of patients with genetically confirmed DNASE1L3 deficiency followed at rheumatology centres in Oman and the UAE. Ethical approval was obtained from all participating centres. Eligible patients had biallelic pathogenic DNASE1L3 variants. SLE classification was assessed using the 2019 American College of Rheumatology/European Alliance of Associations for Rheumatology (ACR/EULAR) criteria⁷ and HUVS according to the Schwartz criteria.⁸ Transition was defined as transfer to adult rheumatology care, occurring at a median age of 13 years in line with regional practice. An additional cross-sectional analysis was performed according to chronological age at last assessment (<18 years versus ≥18 years), independent of transfer to adult services, comparing fulfilment of SLE classification criteria, biopsy-confirmed and proliferative lupus nephritis, disease activity, accrued damage and disease duration between strata.

Data were extracted retrospectively from electronic medical records. Variables included demographics, age at symptom onset and diagnosis, genetic findings, clinical and organ manifestations, laboratory parameters, treatment exposure, hospitalisations, serious infections, disease activity, flares, cumulative damage and disease duration. Organ involvement was based on physician documentation and supporting clinical, laboratory, radiological or histopathological evidence where applicable. Serious infection was defined as infection requiring hospitalisation, intravenous antimicrobials or treatment interruption. Disease activity and flares were assessed longitudinally using SLE Disease Activity Index 2000 (SLEDAI-2K) and predefined flare criteria.⁹ Irreversible organ damage was assessed using the SLICC/ACR Damage Index.¹⁰ Longitudinal outcomes were compared between the last paediatric rheumatology assessment and most recent adult rheumatology follow-up.

Statistical analysis

Descriptive statistics were used to describe the data. For categorical variables, frequencies and percentages were reported. Differences between groups were analysed using Pearson’s χ² tests (or Fisher’s exact tests for expected cells of <5). For continuous variables, mean and SD were used to present the data while analysis was performed using Student’s t-tests. For non-normally distributed continuous variables, median and IQR were used to summarise the data and analysis performed using Wilcoxon-Mann-Whitney tests. An a priori two-tailed level of significance was set at <0.05 level. Statistical analyses were conducted using STATA V.19.0 (STATA Corporation, College Station, Texas, USA). Comparisons between age groups (<18 years versus ≥18 years at last assessment) were unpaired, using appropriate tests.

Results

57 patients from 22 families with homozygous DNASE1L3 variants were enrolled, predominantly from Al Sharqiyah, Oman, with an additional cluster in Al Ain, UAE. Demographics are summarised in table 1. The mean number of affected individuals per family was 2.6. All available parents who underwent genetic testing were heterozygous carriers of the familial variant; however, no heterozygous parent or sibling tested fulfilled classification criteria for SLE or HUVS or had developed clinical autoimmune disease at last follow-up. The cohort included 32 males and 25 females (male-female ratio 1.3–1), with median age at diagnosis of 5 (1.75–13) years and disease duration of 9 (1–22) years. Initial phenotype was SLE in 36 patients (63%) and HUVS in 21 (37%). During follow-up, 56% remained in remission, 39% had a relapsing-remitting course and 5.3% had persistently active disease despite standard treatment. All patients from Oman (n=50) carried the same homozygous DNASE1L3 variant, c.643delT (p.Trp215Glyfs), while all the patients from UAE (n=7) carried the same homozygous variant c.572A&gt;G (p.Asn191Ser).

Table 1. Baseline demographic characteristics of the DNASE1L3 deficiency cohort (n=57).

Characteristic	Value
Total patients, n	57
Sex, n (%)
Male	32 (56.1%)
Female	25 (43.9%)
Male-to-female ratio	1.3:1
Number of families	22
Geographic distribution, n (%)
Al Sharqiyah governorate, Oman	45 (79.0%)
Other Omani regions and UAE	12 (21.0%)
Age at diagnosis, median (IQR), years	5 (1.75–13)
Current age, median (IQR), years	15 (4–32)
Disease duration, median (IQR), years	9 (1–22)
Initial clinical phenotype, n (%)
SLE	36 (63%)
HUVS	21 (37%)
Disease course, n (%)
Persistent remission	32 (56.1%)
Relapsing-remitting	22 (38.6%)
Persistently active	3 (5.3%)
Shared homozygous DNASE1L3 variant	(NM_004944.2): c.643delT (p.Trp215GlyfsX2) (n=50) (NM_004944.4):c.572A>G (p.Asn191Ser) (n=7)

HUVS, hypocomplementaemic urticarial vasculitis syndrome; UAE, United Arab Emirates.

Clinical and laboratory parameters are summarised in table 2. Fulfilment of 2019 ACR/EULAR SLE classification criteria increased from 69% to 84% (p=0.008) during follow-up. Several manifestations declined over time, including malar rash (44% to 17%; p<0.001), urticarial vasculitis (54% to 38%; p=0.035), lymphadenopathy (29% to 11%; p=0.006), and haemolytic anaemia (35% to 17%; p=0.031). By contrast, renal involvement was more frequent on follow-up including; proteinuria (>0.5 g/day) increasing from 33% to 50% (p=0.049), and biopsy-confirmed lupus nephritis increasing from 34% to 46% (p=0.07), though only marginally significant, with proliferative nephritis (class III/IV) predominating at both timepoints. Pulmonary system involvement emerged during adult follow-up but did not reach statistical significance. Hypocomplementaemia persisted in most patients (79% vs 79%; p=1.000).

Table 2. Longitudinal clinical and laboratory characteristics of the DNASE1L3 deficiency cohort.

Parameter n/N (%)	Last paediatric assessment	Last adult follow-up	P value
Classification
Meets 2019 ACR/EULAR SLE criteria	38/55 (69%)	46/55 (84%)	0.008
Meets Schwartz criteria for HUVS	16/55 (29%)	10/55 (18%)	0.109
Cutaneous and constitutional manifestations
Fever	34/54 (63%)	26/54 (48%)	0.185
Malar rash (acute cutaneous)	24/54 (44%)	9/54 (17%)	<0.001
Discoid lesions (chronic cutaneous)	11/54 (20%)	5/54 (9.3%)	0.070
Photosensitivity	12/54 (22%)	10/54 (19%)	0.754
Oral ulcers	4/54 (7.4%)	0/54 (0%)	0.125
Non-scarring alopecia	1/54 (1.9%)	1/54 (1.9%)	1.000
Urticarial vasculitis	30/56 (54%)	21/56 (38%)	0.035
Systemic manifestations
Inflammatory eye disease	22/55 (40%)	22/55 (40%)	1.000
Abdominal pain	17/55 (31%)	12/55 (22%)	0.302
Generalised lymphadenopathy	16/55 (29%)	6/55 (11%)	0.006
Arthritis	28/55 (51%)	24/55 (44%)	0.481
Renal manifestations
Proteinuria (>0.5 g/day)	18/54 (33%)	27/54 (50%)	0.049
Cellular casts	5/53 (9.4%)	7/53 (13%)	0.688
Biopsy-confirmed lupus nephritis	17/50 (34%)	23/50 (46%)	0.070
Serositis	6/54 (11%)	9/54 (17%)	0.508
Pulmonary and neurological manifestations
Pulmonary involvement (haemorrhage)	2/54 (3.7%)	6/54 (11%)	0.289
CNS involvement	2/54 (3.7%)	3/54 (5.6%)	1.000
Haematological manifestations
Haemolytic anaemia	19/54 (35%)	9/54 (17%)	0.031
Leucopenia/lymphopenia	9/54 (17%)	14/54 (26%)	0.179
Thrombocytopenia (<100×10⁹/L)	3/55 (5.5%)	4/55 (7.3%)	1.000
Serology
ANA positive	32/40 (80%)	32/40 (80%)	1.000
Anti-dsDNA positive	13/23 (57%)	8/23 (35%)	0.063
Anti-Smith positive	0/45 (0%)	1/45 (2%)	1.000
Antiphospholipid antibody positive	9/42 (21%)	10/42 (24%)	1.000
ANCA positive	14/32 (44%)	19/32 (59%)	0.125
Low C3 or C4	44/56 (79%)	44/56 (79%)	1.000
Low C3 and C4	36/55 (65%)	35/55 (65%)	1.000
Anti-C1q positive (of those tested)	87%	n/a	n/a
Treatment exposure
Hydroxychloroquine	54/55 (98%)	54/55 (98%)	1.000
Mycophenolate mofetil	35/39 (90%)	36/39 (92%)	1.000
Azathioprine	8/30 (27%)	9/30 (30%)	1.000
Cyclophosphamide	4/24 (17%)	2/24 (8.3%)	0.625
Rituximab	4/25 (16%)	3/25 (12%)	1.000
Methotrexate	3/25 (12%)	1/25 (4.0%)	0.500
Any corticosteroid exposure	49/55 (89%)	29/55 (53%)	<0.001
Current prednisolone dose, mean±SD (mg/day)	n/a	5.9±3.3	—

Denominators vary due to missing data at individual timepoints as indicated.

ACR/EULAR SLE, American College of Rheumatology/European Alliance of Associations for Rheumatology classification criteria for SLE; ANCA, antineutrophil cytoplasmic antibody; anti-dsDNA, anti–double-stranded DNA; CNS, central nervous system; HUVS, hypocomplementaemic urticarial vasculitis syndrome.

Treatment exposure was broadly comparable before and after transition. Hydroxychloroquine use remained unchanged (98% vs 98%; p=1.000), with no significant differences in azathioprine (27% vs 30%; p=1.000), mycophenolate mofetil (90% vs 92%; p=1.000), cyclophosphamide (17% vs 8.3%; p=0.625), methotrexate (12% vs 4.0%; p=0.5) or rituximab exposure (16% vs 12%; p=1.000). Corticosteroid exposure decreased after transition; 89% had prior prednisone exposure, while 53% were receiving current prednisolone at a mean dose of 5.9±3.3 mg/day.

Outcome measures are summarised in table 3. Median SLEDAI-2K scores declined significantly from before transition to last adult follow-up (9 vs 4; p<0.001), but this was not paralleled by improved outcomes. Mean SLICC/ACR Damage Index (SDI) increased from 0.37±0.85 to 0.72±1.17, without statistical significance (p=0.295), and annual flare rates were similar (0.53 vs 0.71; p=0.603). Chronic organ damage increased numerically from 17% to 32%, without statistical significance (p=0.124), with renal, cardiac, pulmonary, neurological and musculoskeletal involvement represented. Osteoporosis and/or vertebral fracture increased from 23% to 41% (p=0.065). Three patients developed end-stage renal disease during adult follow-up, and pulmonary haemorrhage was documented with important prognostic implications.

Table 3. Long-term outcome measures before and after paediatric-to-adult transition.

Outcome measure	Before transition to adult care	Last adult follow-up	P value
Disease activity
SLEDAI-2K score, median (IQR), n=51	9 (3–24)	4 (1–18)	<0.001
Annual flare rate, mean±SD; range	0.53±1.00; 0–3	0.71±1.00; 0–5	0.603
Organ damage accrual
SDI, mean±SD; range	0.37±0.85; 0–3 (n=52)	0.72±1.17; 0–4 (n=37)	0.295
Any chronic organ damage, n/N (%)	9/52 (17%)	11/34 (32%)	0.124*
Osteoporosis and/or vertebral fracture, n/N (%)	12/53 (23%)	14/34 (41%)	0.065
Renal outcomes
End-stage renal disease, n/N (%)	0/57 (0%)	3/37 (8.1%)	0.058
Hospitalisations and infections
Hospital admissions, total/patients; mean±SD; range	49/53; 0.92±1.14; 0–6	27/36; 0.75±0.87; 0–4	0.428
Serious infectious events/year, mean±SD; range	0.40±0.74; 0–3	0.18±0.46; 0–2	0.388
Survival and mortality
Alive, n/N (%)	52/57 (91.2%)	32/37 (86.5%)	0.125
Deceased, n/N (%)	5/57 (8.8%)	5/37 (13.5%)	0.125

^*Denominators vary due to missing data and patients not yet transitioned at time of data collection.

SDI, Systemic Lupus International Collaborating Clinics /American College of Rheumatology Damage Index; SLEDAI-2K, SLE Disease Activity Index 2000; SLICC, Systemic Lupus Interntational Collaborating Clinics.

Mean hospitalisation burden remained comparable (0.92 vs 0.75 per patient; p=0.428) Serious infections per patient-year were numerically lower in adulthood (0.40 vs 0.18; p=0.388), although infective endocarditis and sepsis emerged as key adult diagnoses. Mortality was 8.8% before transition and 13.5% at last adult assessment (p=0.125). Adult deaths were attributed to sepsis with multiorgan failure, infective endocarditis with cardiogenic shock and non-disease-related cause.

Because transfer to adult services occurred at a median age of 13 years, outcomes were additionally examined by chronological age at last assessment. 33 patients (57.9%) were aged <18 years and 24 (42.1%) were aged ≥18 years. All 22 adults with available classification data fulfilled the 2019 ACR/EULAR criteria for SLE (100%), compared with 24 of 33 patients aged <18 years (72.7%; p=0.008). Renal involvement did not differ between the two age groups: biopsy-confirmed lupus nephritis was present in 12 of 23 adults (52.2%) and 15 of 33 patients aged <18 years (45.5%; p=0.786), with proliferative (class III/IV) histology in 43.5% and 42.4%, respectively (p=0.938).

Discussion

This multicentre longitudinal cohort of 57 patients with genetically confirmed DNASE1L3 deficiency, the largest reported to date, demonstrates a clinically important discordance between composite disease activity scores and long-term outcomes. Despite declining SLEDAI-2K scores, patients showed progressive damage accrual, worsening renal involvement, increasing musculoskeletal morbidity and rising mortality during transition into adulthood.

This discordance likely reflects the limitations of SLEDAI in capturing subclinical inflammatory burden in a disease driven by constitutive enzymatic deficiency rather than episodic immune dysregulation.¹¹ Unlike sporadic SLE, DNASE1L3 deficiency causes chronic, genetically determined failure of cell-free DNA clearance. Persistent immune-complex deposition may therefore continue despite fewer overt flares, driving progressive glomerular, pulmonary and vascular injury. The persistence of hypocomplementaemia in our cohort, despite lower SLEDAI-2K scores, suggests ongoing complement activation, consistent with continuing immune-complex activity and supporting DNASE1L3 deficiency as a biologically distinct lupus endotype.

Renal disease remained the dominant manifestation, consistent with previous reports,^{3 5 6} with increasing lupus nephritis despite conventional treatment suggesting incomplete suppression of disease activity in the context of ongoing circulating cell-free DNA. Pulmonary haemorrhage also emerged as an important, potentially under-recognised complication requiring systematic surveillance. In parallel, the high prevalence of osteoporosis likely reflects cumulative corticosteroid exposure during critical periods of bone accrual, indicating that early treatment-related skeletal injury may have long-term consequences despite subsequent dose reduction.

The age-group analysis further supports progressive expansion of the clinical phenotype as shown in online supplemental table 1. Universal fulfilment of SLE classification criteria by adulthood, incomplete in younger patients, is consistent with gradual evolution from an earlier HUVS-predominant presentation to fully expressed SLE. Notably, biopsy-confirmed lupus nephritis, proliferative histology and accrued damage were not significantly greater in adults, indicating that renal injury and damage accrual are already present in childhood and are not confined to adult disease. This suggests that early and sustained control of renal disease is important, and that renal severity is not driven principally by post-pubertal factors, although the modest sample size limits firm conclusions on this point.

The geographic clustering in Al Sharqiyah and Al Ain suggests regional clustering of this pathogenic variant, in the context of familial aggregation and consanguinity reported in these regions.^{3 5 6} The male predominance in our cohort contrasts with the marked female preponderance of sporadic SLE but is consistent with other monogenic lupus syndromes, particularly inherited early complement deficiencies, where sex distribution is more balanced.¹² This supports the concept that high-penetrance genetic defects can drive lupus-like autoimmunity independently of the hormonal factors contributing to female susceptibility in polygenic disease.¹³

Although DNASE1L3 deficiency was initially described in consanguineous Arab families, biallelic pathogenic variants have subsequently been reported in other populations, including European, Turkish, and Chinese families.^14–18 Reported cases consistently support a severe early-onset phenotype spanning HUVS and SLE, with frequent lupus nephritis, antineutrophil cytoplasmic antibody positivity in some patients and occasional pulmonary haemorrhage.^15–18 Recurrent disease-associated variants include c.643delT and c.289_290delAC, while our c.643delT-enriched cohort represents the largest series reported to date and extends the phenotype into long-term follow-up after transition to adult care.¹^4–18

The DNASE1L3 pathway may also be relevant beyond monogenic disease. In sporadic SLE, neutralising autoantibodies with dual reactivity to DNASE1L3 and dsDNA may impair DNASE1L3 activity and are associated with higher disease activity, suggesting that acquired dysfunction may partly mirror inherited deficiency.¹⁹ Similarly, DNASE1L3 knockout mice develop early autoantibodies followed by delayed end-organ damage.¹⁹ Preclinical studies have shown that engineered DNASE1/DNASE1L3 enzyme replacement prevents lupus progression and rescues survival in murine models, supporting its potential as a future therapeutic strategy.²⁰

This study has several limitations. The retrospective design and variable follow-up durations may have introduced ascertainment bias, particularly for documentation-dependent outcomes. Some patients had not yet transitioned to adult services at data collection, so adult data for younger patients reflect early post-transition observations. Transition occurred at a median age of 13 years, younger than in many healthcare systems, which may limit comparability with other childhood-onset SLE cohorts. Loss to follow-up and incomplete historical documentation may have affected SDI scoring. Variable transition protocols across centres may have introduced heterogeneity in post-transition follow-up intensity. The absence of a matched sporadic childhood-onset SLE group limits our ability to determine whether these outcomes are specific to DNASE1L3 deficiency. Although this is the largest reported cohort for this rare condition, statistical power remains limited due to the small sample size.

In conclusion, DNASE1L3 deficiency follows a damage-accruing course into adulthood despite apparent improvement in conventional disease activity measures. These findings challenge reliance on short-term disease activity scores as sufficient indicators of disease control in monogenic lupus and support sustained renal, pulmonary and skeletal surveillance after transition to adult care. Therapeutic strategies that restore defective extracellular chromatin clearance, including enzyme-replacement biologics, warrant evaluation as potential damage-prevention approaches. The findings should be interpreted with caution in light of the small sample size.