Neonatal lupus: Clinical spectrum, biomarkers, pathogenesis, and approach to treatment

Introduction

Fetal exposure to maternal autoantibodies transported across the human placenta via FcγRn can result in a spectrum of organ injury: transient if the tissue has regenerative capacity and permanent if this capacity is absent or limited. An important example of passively acquired autoimmunity in a fetus and neonate is that of neonatal lupus (NL) given the concordance between circulating maternal antibody in the fetus and disease manifestations. NL is often referred to as a syndrome since it can comprise one or more manifestations inclusive of cardiac, cutaneous, liver, and hematologic abnormalities. By definition, the maternal autoantibodies associated with NL are directed to antigen targets within the SSA/Ro–SSB/La ribonucleoprotein complex (60 kDa Ro, 52 kDa Ro, 48 kDa La) (reviewed in Ref. ). NL was coined based on the resemblance of the neonatal rash to the cutaneous lesions seen in SLE. Throughout this chapter, cardiac NL refers to the varied spectrum of cardiac disease, inclusive of complete heart block (CHB), cardiomyopathy, and endocardial fibroelastosis (EFE). The transient manifestations of the syndrome mimic the disease characteristics observed in adolescents or adults (rash and cytopenias). However, the permanent manifestation (heart block), with the exception of one published maternal case, is not observed in the adult despite the presence of identical antibodies in the maternal circulation. The term NL is misleading, as the neonate does not have SLE and often neither does the mother. In many cases, mothers are clinically asymptomatic and only identified to have serologic abnormalities when gestational surveillance reveals fetal bradycardia. Other maternal diseases include an undifferentiated autoimmune syndrome or Sjogren’s syndrome. Despite the rarity of NL, the syndrome continues to be investigated with advances from the bench and bedside, which have been incorporated in this revised chapter.

Risk of cardiac NL and population prevalence

The risk of having a child with cardiac NL is approximately 2% for an antiSSA/Ro-positive woman who has either never been pregnant or has previously had only healthy offspring. If an antiSSA/Ro-positive mother has a previous child affected with cardiac NL or cutaneous NL, the risk of recurrence is 18% and 13%, respectively. In general, women with low titers of antiSSA/Ro antibodies are at less risk than those with high titer antibodies ; however, there is considerable overlap in antibody levels between affected and unaffected cases and most women have high titers that remain stable over time. The risk of developing cutaneous manifestations of NL is 7%–16%, and the recurrence rate of cutaneous NL is estimated to be between 23% and 29%.

The population prevalence of cardiac NL in Finland was reported at 1:17,000 live births with the highest annual estimates at 1:6,500. However, this may be an underestimation since only children with pacemakers were included and fetal deaths were not captured. In children born with cardiac NL in the absence of documented structural abnormalities, antiRo antibodies are found in over 85%.

The prevalence of antiSSA/Ro antibodies had been initially reported at 0.20%–0.72% in female blood donors, 0.87% in pregnant women, and more recently at 0.86% in healthy females in the general population. The prevalence reported in the latter study may be even higher since only those positive for antinuclear antibodies (which may not always detect antiSSA/Ro) were then tested for antiSSA/Ro antibodies. For patients with SLE the prevalence of this antibody reactivity is estimated at 40% and in those with Sjogren’s syndrome between 60% and 100%.

In aggregate, if the true prevalence of antiSSA/Ro approaches 0.9% and cardiac NL occurs in 2% and recurrence in 18%, this could yield approximately 600–700 cases per year based on the 2017 National Vital Statistics of 3,855,500 births. Thus thousands of women in the United States may be faced with the risk of cardiac NL in their offspring, yet prenatal testing does not include an evaluation of antiSSA/Ro antibodies.

Transient clinical manifestations of NL: cutaneous, hepatic, hematologic, and neurologic

In contrast to the in utero detection of cardiac NL, cutaneous disease most often appears after birth, with a mean time of detection at 6 weeks and mean duration of 17 weeks. Albeit the specificity of the maternal autoantibodies may be identical, the discordant timing of the cardiac and cutaneous disease supports distinct initiators or amplifiers of injury. The strong association of cutaneous lesions with UV exposure suggests that apoptosis of the keratinocytes and surface translocation of SSA/Ro may result in immune complex formation and tissue injury. All mothers with antiSSA/Ro antibodies should be counseled regarding UV protection of their infants as both a preventative and therapeutic measure.

The rash is characterized by erythematous annular lesions or arcuate macules with slight central atrophy and raised active margins, which are located primarily on the scalp and face with a characteristic predilection for the upper eyelids. A raccoon-like appearance should immediately raise suspicion for NL. A review of the corporeal distribution of rash among 57 infants with cutaneous NL enrolled in the Research Registry for Neonatal Lupus (RRNL) revealed that 100% had facial involvement. Other affected areas included the scalp, trunk, extremities, neck, intertriginous areas, and rarely the palms or soles, in descending order. The NL rash resembles subacute cutaneous lupus erythematosus, with basal cell damage in the epidermis and a superficial monocyte cell infiltrate in the upper dermis. Immunofluorescence staining of skin biopsies reveals IgG deposition within the epidermis. Of relevance to the pathogenesis of tissue injury mediated by antiSSA/Ro antibodies, histiocytes consistent with M2 macrophages have been identified in the skin lesion of a neonate with cutaneous NL. Although purely speculative, this observation echoes the macrophage infiltrations noted in cardiac NL, suggesting a potential link between these two manifestations: one induced by UV light and the other by apoptosis (see the following), both exposing intracellular antigen and inciting an inflammatory infiltrate in response to immune complexes.

The rash is usually self-limiting and almost always resolves by approximately 8 months of age, coincident with the clearance of maternal antibodies from the child’s circulation. Residual skin abnormalities are uncommon but can include atrophy, scarring, pitting, hypopigmentation, or hyperpigmentation, and telangiectasias. On rare occasion, neither antiSSA/Ro nor SSB/La is detected in the maternal sera of a child with classic NL lesions, but rather antibodies to another ribonucleoprotein, U1-RNP, are present. Thus the evaluation of this latter reactivity should be considered when facing the differential diagnosis of NL in a child with characteristic rash and maternal disease is unsuspected.

In general the self-limiting nature of the rash precludes therapy. Topical steroids (nonfluorinated) have been used. However, data from the RRNL revealed no significant differences in outcome with or without treatment. Systemic therapies are not recommended.

Neonatal liver disease is associated with maternal antiSSA/Ro but its true prevalence is unknown since routine testing at birth does not include a liver enzyme profile. In one prospective study, 26% of children born to mothers with antiSSA/Ro had elevated liver enzymes. Laxer described NL associated with significant hepatic involvement in four infants, three living and one who died postnatally. The clinical picture in these neonates was cholestatic. Pathologic changes included giant cell transformation, ductal obstruction, and extramedullary hematopoiesis. Lee and coworkers investigated the incidence of hepatobiliary manifestations among 219 NL patients in the RRNL and noted that recognized hepatobiliary disease occurred in 19 (9%) of 219 infants, usually in conjunction with either cardiac or cutaneous involvement. Three clinical variants were observed: (1) severe liver failure present during gestation or in the neonatal period (least common); (2) conjugated hyperbilirubinemia with mild or no elevations of aminotransferases occurring in the first few weeks of life; and (3) mild elevations of aminotransferases occurring at approximately 2–3 months of life. The prognosis for the children in the last two categories was excellent.

Hematologic manifestations of NL include thrombocytopenia, neutropenia, anemia, and, very rarely, aplastic anemia. Thrombocytopenia was present in 10% of the neonates referred to Lee and coworkers. While NL thrombocytopenia is presumed to be autoimmune in nature, its exact pathogenesis remains unclear since it is uncertain whether antiplatelet-specific antibodies or antiSSA/Ro-SSB/La antibodies target the surface of fetal platelets. With regard to the pathogenesis of neutropenia, in vitro exposure of intact neutrophils to antiSSA/Ro-positive maternal and/or infant serum from affected families results in immunoglobulin deposition, suggesting a possible immune-mediated basis for NL neutropenia. Indeed, the neutrophil/immunoglobulin interactions were neutralized by preincubating the sera with 60 kDa Ro antigen that bound the autoantibody, suggesting that anti60 kDa SSA/Ro directly drives the pathogenesis of neutropenia. In one prospective study, 25 of 107 infants born to mothers with antiSSA/Ro or antiSSB/La antibodies had neutropenia but no cases of neonatal sepsis occurred.

Neurologic dysfunction has been reported in offspring of mothers with antiSSA/Ro antibodies. In a Canadian cohort of 87 infants exposed to maternal antiSSA/Ro and/or antiSSB/La antibodies, 8% (5/47 with and 2/40 without another manifestation of NL) had hydrocephalus, all but one resolving spontaneously. Maternal immunological dysfunction has been associated with reports of developmental language delay, learning difficulties, and left handedness.

In a retrospective study based on detailed questionnaires, telephone interviews, and reviews of medical records of children with NL in the RRNL, their unaffected siblings, and healthy friend controls, it was noted that behavioral problems, either isolated or associated with attention disorder, were present in all groups with no statistical difference. The prevalence of depression, anxiety, developmental delays, learning disability, hearing and speech problems, and use of stimulants were also not significantly different between groups. The authors suggested that parental reporting of neuropsychiatric abnormalities was high in antibody-exposed children; however, it did not meet statistical significance when compared to the controls. More recently, in a study from Sweden, impaired neurodevelopment was reported in 16% of antiRo-exposed children [60 siblings with and 54 without CHB (18/114)]. Reported problems included speech (9%), motor (8%) and learning (8%) impairment, attention deficit (5%), and behavioral impairment (4%). Impairment in motor skill development was more common in boys ( P < .001) if the child was born preterm ( P < .001). Learning impairment was significantly influenced by maternal SLE ( P < .005), while attention deficits were influenced by both maternal SLE ( P < .05) and cardiac NL in the child ( P < .05).