Current insights into K-associated fetal anemia and potential treatment strategies for sensitized pregnancies

Abstract

K-associated anemic disease of the fetus and newborn (K-ADFN) is a rare but life-threatening disease in which maternal alloantibodies cross the placenta and can mediate an immune attack on fetal red blood cells expressing the K antigen. A considerably more common disease, D-associated hemolytic disease of the fetus and newborn (D-HDFN), can be prophylactically treated using polyclonal α-D antibody preparations. Currently, no such prophylactic treatment exists for K-associated fetal anemia, and disease is usually treated with intrauterine blood transfusions. Here we review current understanding of the biology of K-associated fetal anemia, how the maternal immune system is sensitized to fetal red blood cells, and what is understood about potential mechanisms of prophylactic HDFN interventions. Given the apparent challenges associated with preventing alloimmunization, we highlight novel strategies for treating sensitized mothers to prevent fetal anemia that may hold promise not only for K-mediated disease, but also for other pathogenic alloantibody responses.

1. Introduction

Hemolytic disease of the fetus and newborn (HDFN) can be a life-threatening disease caused by maternal alloantibodies specific for non-self red blood cell (RBC) antigens that can mediate destruction of fetal RBC and/or erythroid precursors. The resulting reduction in fetal RBC may lead to fetal anemia, hydrops, and death¹. HDFN occurs in the United States at a rate of 1695/100,000 newborns². Prophylaxis and treatment options for these cases are limited and vary with the type of RBC antigen mismatch^3,4. Therefore, understanding the underlying biology of each RBC alloantigen has the potential to contribute to identifying pregnancies at highest risk for this complication, detecting and treating disease, and supporting discovery of new prophylactic or therapeutic interventions.

In most cases, sensitization of the maternal immune system to RBC alloantigens occurs before the pregnancy that is affected by fetal anemia⁴. Exposure to alloantigen occurs either by transfusion of blood products containing mismatched RBC antigens or during the natural breakdown of placenta or hemorrhaging that leaks fetal blood into the mother’s circulation in the course of a previous pregnancy (Figure 1)⁵. This sensitization is most commonly caused by antigens belonging to six blood group systems: Rhesus (Rh), ABO, Kell, Duffy, Kidd, and MNS (Figure 2)^6,7. The severity of fetal anemia varies with antigen for many reasons, including onset of antigen expression during fetal development, timing and prevalence of antigen expression during RBC development, immunogenicity of the antigen, and characteristics of the maternal antibody response⁸. While α-Rh (anti-Rh antibody), particularly α-D is most common, α-K is the second most common cause of severe fetal anemia⁹. Expression differences between Kell and Rh antigens mean that fetal anemia mediated by the two blood systems progresses differently, as α-K antibodies have the added ability to suppress erythrocyte precursors¹⁰. Unlike Rh-disease, Kell-associated disease can thus present with severe fetal anemia without features associated with hemolysis of mature RBCs¹¹. As a result, some have advocated for the broader definition of α-K disease as Kell-associated anemic disease of the fetus and newborn (K-ADFN)¹².

Figure 1: Maternal K sensitization and α-K antibody mechanism of action.

A. Mothers are usually sensitized either by a transfusion with K blood or by hemorrhage during previous pregnancy. B. K mismatch or risk of mismatch can be identified with labwork and noninvasive prenatal testing. C. To cause fetal anemia, α-K IgG must cross the placenta via FcRn.

Figure 2: Blood Group Antigens Associated with Maternal Sensitization and Fetal Anemia.

Left: Schematic of common RBC blood group antigens associated with HDFN. Right: Antigen classes, allotypic variants and their frequency in different populations, and the degree to which they are associated with severe fetal anemia^4,6–8.

Recently, there have been advances both in our understanding of alloimmunization and in the treatment of women sensitized to different antigens that have the potential to improve the prevention and management of alloimmune fetal anemia. Since the 1960’s, the prevalence of α-D-associated HDFN (D-HDFN) has decreased due to successful prevention of maternal sensitization with polyclonal α-D immunoglobulin (RhIg)¹³. No such treatment exists for K-associated fetal anemia, which is rare but often serious when it occurs. As an example, in 1986, a 16 year retrospective study of 127,076 American patients by Caine et al., found K sensitization in 127 pregnancies (0.1%)¹⁴. Of the 13 pregnancies with maternal sensitization to fetal antigen, five pregnancies (38%) had severe outcomes: two neonates suffered hydrops fetalis and three died in utero^9,14. Current treatment for K-associated disease is to treat the mother with IVIG and to manage fetal anemia through intrauterine blood transfusions, suggesting the value of more research into means to reduce either or both the risk and consequences of maternal alloimmunization. This review will focus on recent work investigating the challenges in developing prophylactic recombinant Ab therapeutics to prevent ADFN/HDFN and the development of new treatment strategies to support sensitized pregnancies. We will highlight work related to K-ADFN as it is a common form of severe disease, but many of the mechanistic lessons and treatment strategies may apply across sensitization to a variety of antigens.

2. Characteristics of the K antigen and current treatment of K-ADFN

A diversity of polymorphic surface proteins are expressed on RBC. Among other factors, these alloantigens cause HDFN with differing prevalence and consequence due to variable characteristics of population genetics and alloantigen biology.

2.1. KEL1 antigen and immunogenicity

Maternal alloantibodies that mediate K-ADFN are raised against allotypes of a membrane glycoprotein expressed early in RBC development^15–17. Most often, K-ADFN is associated with the KEL1 allele, in which a single nucleotide polymorphism leads to a substitution of a highly conserved threonine with methionine at amino acid 193 in the Kell protein (Figure 3). This mutation disrupts a glycosylation site, such that an N-linked glycosylation sequon present in the K negative phenotype is absent in the K phenotype, resulting in expression of an immunogenic epitope^15,18,19. If a mother has been previously exposed to K antigen, she may produce α-K immunoglobulin G (IgG) that then has the potential to mediate destruction of fetal RBCs when transported across the placenta to the fetus by the neonatal Fc receptor (FcRn)²⁰, which has important implications for novel approaches to treatment. Based on the incidence of K-sensitized, antigen-mismatched pregnancies reported in retrospective studies¹⁴ as well as global annual pregnancies rates²¹, we estimate about 10,000 pregnancies are affected by K-ADFN worldwide per year.

Figure 3. Amino Acid substitution in K antigen arising from a mutation in the KEL1 gene.

An amino acid polymorphism that changes a threonine to methionine eliminates an N-linked glycosylation motif, resulting in presentation of an antigenic epitope.

2.2. Current treatment of K-ADFN

Early detection of K-mediated fetal anemia is essential for positive outcomes. Thirty years ago, a study of 208 pregnancies reported that 89% of fetuses expressing D antigen survived using intrauterine transfusion as treatment for fetal anemia²². The same study reported that only 56% of those expressing K survived, likely in part because of delayed recognition and referral for intrauterine transfusion²². Because K is expressed in primitive erythrocytes, maternal α-K alloantibodies inhibit erythroid progenitor cell growth^10,11,23,24. In vitro assays comparing the effect of α-D or α-K IgG monoclonal antibodies on progenitor cell proliferation found α-K IgG specifically inhibited growth of K-expressing erythroid progenitor cells and had no effect on the proliferation of either K negative erythroid progenitor cells or progenitor cells of other lineages¹⁰. Perhaps because of this early expression of the antigen, there is much less hemolysis in K-associated fetal anemia than in the more common Rh-associated HDFN¹², making it challenging to diagnose. Historically, the severity of fetal hemolysis in Rh-associated disease can be determined with spectrophotometric measurement of bilirubin in amniotic fluid, but in the case of α-K-mediated disease, bilirubin concentration is low, and diagnosis is less accurate for detecting severe K-associated anemia (83–89% accurate) than for Rh-associated disease (95%)^9,25. Cordocentesis-based measurement of fetal hemoglobin can provide a better indicator of severity of K-associated fetal anemia²⁶. Because of the invasiveness of both amniocentesis and cordocentesis, Doppler ultrasonography measurements of fetal middle cerebral artery peak systolic velocity have been correlated to cordocentesis measurements^27–29. These ultrasonography based measurements allow for noninvasive monitoring of fetal anemia mediated by many antigens^27–29. In the case of a K-sensitized mother with potential for a K fetus, the most effective course is to monitor the fetus for signs of anemia and hydrops every 1–2 weeks by ultrasound starting at 17 weeks³⁰.

When identified, fetal anemia is typically treated with intrauterine blood transfusions (IUT). IUT is effective but carries a rate of fetal loss of 8.5% before 20 weeks estimated gestational age (EGA), and 0.9% after 20 weeks³¹. In K-associated disease, maternal antibody titer does not significantly correlate well with the severity of fetal anemia or the EGA of the first necessary IUT³⁰. This lack of correlation suggests that qualitative aspects of the antibody response may have an important influence on its pathogenicity^32,33, and results in conflicting practices concerning the influence of antibody titers in guiding clinical response³³. Fortunately, for children who survive with IUT treatment, cognitive and physical outcomes are good³⁴.

3. Recent work to define the mechanisms of sensitization to RBC antigens in mouse models

Recent studies in preclinical models have shed light on processes involved in alloimmunization^35,36. An important fundamental consideration in understanding alloimmunization is to identify the immune cells that first recognize the antigenic RBC. A recent study in mice indicated that upon transfusion, RBC expressing a model antigen comprised of hen lysozyme, ovalbumin, and human Duffy antigen (HOD), colocalized with marginal zone (MZ) B cells³⁷. In mice depleted of MZ B cells, lower titers of α-HOD antibody (IgM and IgG) were detected, whereas depletion of follicular B cells had no effect on α-HOD titer³⁷. This identification of MZ B cells as an important cell type in the process of alloimmunization suggests a new potential target for prophylactic treatment. This model also points to a role for CD4+ T cells in sensitization, as mice depleted of CD4+ T cells did not raise α-HOD responses. In addition, α-HOD responses were durable, lasting for more than three months, suggesting CD4+ T cell-driven differentiation of B cells to antibody secreting cells³⁷. An analogous study that focused on Kell sensitization in mice uncovered similar MZ B cell-dependent sensitization responses, though these were independent of CD4⁺ T cells, again highlighting that the mechanism of alloimmune responses may vary by antigen^38,39.

Other host factors also play a role. In a model of K sensitization in which healthy mice can be protected from sensitization with α-K prophylactic and K⁺ RBC³⁸, researchers observed increased rates of sensitization when they also induced viral-like inflammation in mice by intraperitoneal administration of the adjuvant poly (I:C)⁴⁰. This sensitization was dependent on the complement pathway, as mice lacking complement pathway component proteins C3 or C1q did not show increased sensitization⁴⁰. In in vitro studies, B cells isolated from complement knock-out mice were less likely to bind complement-fixed, opsonized RBC compared to B cells from wild-type mice⁴⁰. This study highlights the potential roles of inflammation state and complement in mediating B cell activation and maternal sensitization to antigenic RBCs, suggesting further targets that may be relevant to strategies to prevent alloimmunization.

Overall, while these and other preclinical studies modeling RBC sensitization vary in biological relevance, they are consistent in identifying that multiple factors play a role in the likelihood and characteristics of humoral alloimmunity.

4. Developing new treatments for HDFN

To review new strategies for treatment of HDFN, we consider the prophylaxis of HDFN by RhIg (e.g. RhoGAM), which has changed the landscape of D-mediated disease. Interest in the mechanism of RhIg has also motivated work to better understand alloimmunization more broadly^41–46. The success of RhIg is tempered by challenges in reproducing the effect with monoclonal antibodies (mAbs). These challenges mean that while expanding prophylactic treatment to HDFN mediated by other antigens is of high interest, this goal may not be achievable. Because antigen-specific prophylactics are likely to be difficult to develop, alternative treatments to better manage pregnancies in women who are already sensitized that are agnostic to the RBC antigen involved would be of high clinical value.

4.1. Prophylaxis of HDFN

In principle, HDFN can not only be treated to reduce disease severity and sequelae, but also can be prevented by interventions that reduce the likelihood of alloimmunization. This strategy has been enormously successful in the case of D-HDFN. Worldwide, D-HDFN is estimated to occur in around 276 out of 100,000 pregnancies⁴⁷. Of those that develop D-HDFN and go untreated, 50% will either die or develop brain damage as a result^47,48. The introduction of RhIg in 1968 led to the decrease of D-associated alloimmunization in pregnancies by 95%^49,50. In a Dutch study, Koelewijn and colleagues reported antenatal RhIg treatment lowered the frequency of pregnancies in D negative mothers resulting in Rh sensitization from 0.66% in pregnancies given only postpartum RhIg (200 μg dose) to 0.31% in pregnancies given both antepartum and postpartum RhIg (both 200 μg doses)⁵¹, a rate of sensitization that represents a 50-fold decrease from the risk of sensitization without any treatment⁵². A corresponding halving of the rate of severe HDFN was also seen in mothers given antenatal RhIg⁵¹. Generally, RhIg treatment is safe, effective, and specific, successfully reducing the frequency of D sensitization with minimal risk of side effects^53,54. Successful prevention of D-HDFN using RhIg is a major achievement in the fields of fetal care and immunotherapy. The efficacy of this treatment might guide the development of new treatments for K-mediated disease^53,55.

By contrast, the shortcomings of RhIg - the difficulty of manufacture, the mysterious mechanism of action, and the challenges of developing a recombinant version of the therapy – temper the feasibility of identifying antigen-specific prophylactics. RhIg is pooled, polyclonal IgG antibody from D-negative donors sensitized to D-positive RBC⁵⁶. Therefore, RhIg is limited in supply, variable in nature, somewhat expensive to produce, and carries a small risk of transmission of blood borne disease⁵⁶. The exact dosing of RhIg varies by country and case, but most women receive 2–3 doses between 100–300 μg RhIg⁵¹. Globally, about 2.7 million women are treated with RhIg per year⁵⁷. Still, much of the need in poorer countries remains unmet, and recent work suggests a gap in needed D prophylaxis of about 2.9 million doses⁵⁷. Therefore, the development of prophylatic α-D monoclonal antibody (mAb) treatments that could be mass-produced to replace RhIg is of significant interest. Despite the prevalence of mAb therapies in diverse disease settings and the long-standing, widespread use of polyclonal RhIg, a monoclonal form of RhIg (RhoClone^®), has only recently been approved and is available in relatively few countries⁵⁸. Lessons from efforts to develop prophylactic α-D mAbs illustrate the complexity of developing mAb therapeutics in the context of prevention of alloimmunization.

4.2. Challenges in developing mAbs to prevent sensitization

Until the early 2000’s, clearance of antigenic RBC was a favored hypothesis for the efficacy of RhIg, as clearance of antigenic fetal red blood cells from maternal circulation is an effective metric for ensuring the correct dosing of RhIg^59,60. Clearance of RBC was used as a surrogate measure to identify promising candidate monoclonal D-specific antibodies, but the majority of these mAbs had disappointingly variable effects on sensitization. Candidates with varying IgG subclasses and profiles for recognition of antigen, lysis or phagocytosis of antigen-expressing cells, and clearance of RBC in vivo have been considered^61,62, and recent work has sought to clone novel candidates from RhIg donors⁶³. In the late 1990’s and early 2000’s, at least 15 first-in-human trials of mAb and recombinant Ab RhIg replacement candidates were completed⁵⁰. In addition to safety, these studies assessed the ability of treatments to clear antigenic RBC, and in some cases, the ability to reduce the probability of alloimmunization⁵⁰. Although candidates were initially identified through similar performance in model clearance assays, a wide variety of outcomes were seen in the clinic⁵⁰. Sadly, one proposed mAb treatment, chosen for rapid RBC clearance, led to enhanced alloimmunization in a small human clinical trial^50,59. Other treatments had low bioavailability and variable in vivo effects. Perhaps most intriguingly, performance of two mAbs, BRAD-3 and BRAD-5, changed when they were manufactured using a human B-lymphoblastoid cell line or Chinese Hamster Ovary cells, leading to the hypothesis that antibody glycosylation patterns, which vary by expression host and affect binding to FcγR, were essential to mAb activity⁵⁰.

The importance of mAb glycosylation has been demonstrated in a few contexts. Broadly, galactosylated and asialylated IgG’s have been shown to be proinflammatory and to increase ADCC activity through FcγRIIIa interactions⁶⁴, and increased levels of IgG galactosylation and sialylation are associated with onset of rheumatoid arthritis remission during pregnancy⁶⁵. Characterization of 23 α-D mAbs, many investigated clinically, demonstrated a diversity of glycosylation states⁶⁶. Across this panel of mAbs, differences in fucose and galactose content resulted in differences in FcγRIIIa binding and activity in a cell-based antibody-dependent cellular cytotoxicity assay (ADCC), as well as in in vivo clearance assays in humans⁶⁶. In another study to probe the importance of glycosylation state to prophylaxis, D-specific antibodies were purified from a polyclonal RhIg treatment and glycoprofiled⁶⁷. The D-specific antibody fraction was more highly galactosylated and sialylated than the total IgG found in an RhIg product, indicating that immunomodulatory and lower ADCC activity may be preferable for prophylaxis⁶⁷. This increase in galactosylation of D-specific IgG was also seen in a survey of RhIg from different manufacturers⁶⁸. Glycosylation patterns of α-K, particularly the fucose content of the Fc domain glycans, have been shown to correlate with severity of disease in K-ADFN^69,70. Similar observations have been made in α-D⁷¹, α-c HDFN⁷⁰, and thrombocytopenia⁷². These glycosylation patterns are stable through the pregnancy suggesting that, in addition to being an important factor in the design of prophylactic mAbs, glycosylation patterns of α-antigen antibodies could be leveraged in developing treatment plans for affected pregnancies^69,72.

Work to extend prophylactics to K sensitization in animal models has hinted at similarly complex responses to prophylactic mAbs and the weakness of RBC clearance as a surrogate measure of efficacy. The importance of prophylactic mAb subclass to sensitization, but not RBC clearance, was systematically demonstrated in a K antigen mouse model of alloimmunization. While this is a useful model in the study of sensitization, an important caveat is that recipient mice lack the Kell protein expression altogether and are then transfused with K⁺ RBC, whereas in human disease, mother and baby express two allotypes of the same protein (Figure 3). A strength of this study is that mice were treated with antibodies of identical specificity to a single Kell epitope and recombinantly expressed in the complete panel of mouse IgG subclasses (IgG1, IgG2a, IgG2b, IgG2c, and IgG3), which exhibit variable effector function activities. The mice were then sensitized with mouse RBCs engineered to express Kell. Each subclass-switched mAb treatment led to clearance of K⁺ RBCs, but mAb-treated mice raised differential titers of α-K antibody relative to those treated with a buffer control⁷³. Subclasses with high affinity for FcγR tended to enhance alloimmunization, and those with decreased effector function generally decreased alloimmunization compared to buffer control⁷³. IgG subclass and FcγR-dependent sensitization was also seen in a second model in which mice were passively immunized with α-Duffy IgG, and then exposed to RBCs expressing a fusion protein (HOD) including part of the human Duffy antigen⁷⁴. In this study, sensitization in the context of Duffy-specific mAb was also associated with epitope spreading—the induction of antibodies toward other, co-expressed RBC antigens⁷⁴. Beyond supporting the finding that subclass is important to modulating immune response, experiments here identify CD8⁺ CD11b⁺ dendritic cells (DC) as essential to antigenic RBC clearance and show that this clearance coincides with increased CD4⁺ T cell proliferation⁷⁴. Hoping to isolate RBC clearance and sensitization to a particular FcγR and immune compartment, an ambitious series of conditional FcR γ chain knockout mice strains were created to attempt to modulate FcγR activity in subsets of immune cells. Treatment and sensitization of mice with variation in FcγR expression successfully decoupled RBC clearance from alloantibody response in this model⁷⁴, and made apparent that the phagocytic cells that clear antigenic RBCs are distinct from the antigen presenting cells that drive sensitization leading to increased alloantibody titers⁷⁴. In a follow up study, either genetic knockout or mAb blocking of murine FcγRIV prevented sensitization and RBC clearance by IgG2c subclass α-HOD⁷⁵. This finding is consistent with the identification of the role of DCs in clearance as in mice FcγRIV is expressed on splenic DCs⁷⁶.

In a similar K antigen mouse model, Stowell and coworkers proposed modulation of antigen expression rather than RBC clearance as the mechanism of antibody prophylaxis to alloimmunization³⁸. They observed that alloimmunization could be prevented when mice transfused with mouse RBCs expressing K were treated with α-K sera or polyclonal α-K IgG, alloimmunization could be prevented. Mice were not protected when sera or polyclonal antibody specificity were mismatched with transfused RBC antigen. Intriguingly, by monitoring transfused RBC using either a tracking dye or GFP expression rather than antigen expression, about 50% of the transfused RBCs remained in circulation for 20 days post transfusion, but detection of the K antigen expression was nearly entirely diminished within 24 hours of antibody treatment. In contrast, K antigen remained detectable over the same time period in mice transfused with K RBC alone and was stable for weeks in agammaglobulinemic mice³⁸. In a recent study, antibodies that lead to the removal of HOD antigen from RBC suppressed sensitization, confirming antigen loss rather than RBC clearance as an important mechanism in tolerance⁷⁷. Together these studies point to the importance of α-antigen mAb effector function in sensitization and have uncovered new insight into the effects of antigen dose and antigen loss on RBCs, although care should be taken in generalizing lessons across epitopes and from mouse models to the human immune system.

Overall, results from prophylactic mAb studies in animal and human experiments certainly reinforce the complexity of the biology of alloimmunization, as well as the likelihood that multiple mechanisms are at play in sensitization to RBC antigens. Nonetheless, they may also suggest some generalizable concepts. Namely, that clearance of antigen, whether by antigenic modulation, RBC clearance, or other mechanisms is desirable, but outcomes may be best if this clearance is achieved in the absence of inflammation, such as may be induced by non-cytolytic mAbs. Additional effects on B and T cell immunomodulation, including changes that may be associated with pregnancy, appear difficult to model, but may be important factors clinically.

5. Antigen-independent treatments

Beyond the open questions in understanding therapies in the absence of clear mechanism of action, typing the mother-infant dyad can be difficult, and the occurrence of fetal anemia caused by antigens other than D is rare. Therefore, there is a clinical need for alternative interventions that could prevent or treat fetal anemia across different alloantigens in the context of existing sensitization.

5.1. Motivation, benefits, and risks of antigen-independent treatment

Preventing maternal sensitization requires blood group typing of mothers, and potentially genotyping of the father and fetus. Noninvasive prenatal testing, in which fetal (cell-free) DNA is sequenced from a maternal blood sample, has improved screening for D+ fetuses and emerging technologies may make it possible to screen for other blood antigens including Kell blood group antigens⁷⁸. Given the rarity of K-mediated fetal anemia, even if effective strategies to prevent K alloimmunization were available, implementation of testing would likely face challenges to its cost-effectiveness.

In contrast, novel approaches to prevent anemia after sensitization have the potential to improve outcomes. Additionally, these strategies could potentially be employed for diverse cases of fetal anemia, independent of the identity of the specific blood antigen mismatch. These strategies include removing maternal alloantibodies from circulation, blocking alloantibody transfer across the placenta, blocking alloantibody and RBC antigen interactions, and reducing maternal antibody production^37,79. A downside to these treatments is that they also affect the transfer of most or all maternal antibodies, which is expected to leave the neonate more vulnerable to infectious disease. The importance of maternal-fetal antibody transfer to health in early life has been demonstrated in both the protection of neonates through maternal vaccination⁸⁰ and in the increased risk of neonatal infection when maternal antibody transfer is impaired⁸¹.

5.2. Blocking alloantibody transfer

Because maternal alloantibodies must be transferred to the fetus via the placenta to cause ADFN/HDFN, manipulating maternal-fetal antibody transfer may lead to generalizable, or antigen-independent treatment modalities. All molecules passing from maternal to fetal blood must cross this barrier, which is made up of two layers of cells, the first being multinucleated syncytiotrophoblasts and the second the endothelial cells of fetal capillaries²⁰. Molecules less than 500 Da can usually passively diffuse across the placenta, while molecules of higher molecular weight generally do not transfer at all²⁰. IgG antibodies, with molecular weight around 160 kDa, are an exception to this rule and are the sole antibody isotype to be transferred across the placenta in significant amounts²⁰. IgG transfer occurs via pH-dependent binding of IgG to the neonatal Fc receptor (FcRn), which is an integral membrane glycoprotein that supports both the long plasma half-life of IgG and its transport across the placenta from acidifying endosomes^82,83. Though transfer efficiency varies by IgG subclass, IgG is actively transported across the placenta⁸⁴. Therefore, when the mother’s blood exhibits high concentrations of IgG such as α-K, these antibodies will readily be transported through the placenta and into the fetal blood where they may have detrimental effects, suggesting the value of interventions that would restrict placental transfer of antibodies⁸⁵.

One such FcRn blocking strategy is treatment with nipocalimab (also called M281), which is an aglycosylated recombinant human IgG1 mAb with an affinity for FcRn 1,000 times that of endogenous IgG Fc⁷⁹, developed with treatment of both auto- and alloimmune antibody-dependent diseases as potential indications. Preliminary studies from animal models, human placenta studies, and human trials suggest blocking FcRn with this antibody drug potentially protects the fetus in two ways^79,86–89. First, this intervention prevents maternal IgG from crossing the placenta^86,88, and second, it inhibits the IgG recycling function of FcRn, thereby decreasing the circulating concentration of maternal IgG, including pathogenic IgG alloantibodies^79,89. Phase I trials of nipocalimab proved that it was safe in humans and did not increase the risk of infection⁹⁰. In June of 2023, results from UNITY (NCT03842189), a phase II clinical trial to assess the ability of nipocalimab to result in live birth at 32 weeks of gestation or later without the need for IUT in alloimmunized pregnancies at high risk of early-onset severe HDFN were announced. Whereas only 10% of reference pregnancies met this outcome, 54% of patients who received nipocalimab did, with a mean gestational age greater than 37 weeks⁹¹. While inclusion criteria included α-D titers ≥32 or α-K titers ≥4, it is not yet clear how many participants exhibited which alloreactivity. While promising, this study relied on published and unpublished data as a historical reference for comparison to outcomes with the 13 subjects enrolled. A more rigorous randomized and placebo controlled Phase III study in 120 participants is planned (NCT05912517). Other FcRn blocking mAbs, such as efgartigimod⁹² and rozanolixizumab⁹³, which have been approved for treatment of generalized myasthenia gravis based on reducing circulating antibody and are in development for treatment of disease mediated by other autoantibodies, may hold similar promise in preventing or treating ADFN/HDFN^94,95.

5.3. Reducing levels of maternal alloantibody

Blocking FcRn is not the only means to reduce the levels of pathogenic maternal alloantibodies. In fact, in a small case study, two cases of severe K-mediated disease were successfully treated with three such strategies: immunoadsorption, intravenous immunoglobulins (IVIG), and rituximab. Treatment was started at 14 and 21 weeks EGA, at which point the mothers had α-K antibody titers of 2000 and 8000 respectively⁹⁶. Each component of this treatment mitigates the severity of fetal anemia. First, immunoadsorption is an apheresis technique that selectively removes immunoglobulins and immune complexes from the blood. In the context of HDFN, the goal of this intervention would be to lower the titer of pathogenic IgG by removal of IgG from circulation⁷⁹. This process removes beneficial antibodies as well, and requires highly skilled technicians and specialized equipment, making it unlikely to be widely adopted⁷⁹. Second, though used in many indications, the mechanism of action of IVIG is unclear, but in this context, it is thought to dilute maternal IgG and to occupy FcRn binding sites. This occupancy is thought to result in reduced transfer of endogenous maternal IgG to the fetus, and/or to reduce antibody expression by plasmablasts through a rheostat-style mechanism³¹. A fraction of antibodies included in IVIG, likely those with Fc sialylation, may also be immunomodulatory⁹⁷. Studies of both RhIg and IVIG point to antigen-independent action of these treatments that modulate cytokine expression likely impacting T and B cell activity^97–101. These antibody Fc domain-based interactions are likely context-dependent and the receptors involved in signaling in humans have yet to be identified¹⁰². Lastly, rituximab mediates CD20+ B cell killing, and significantly reduces the number of circulating plasmablasts, resulting in temporary reduction in α-K^96,103,104. This combination of treatments lowered the α-K antibody titers to 32 in both mothers⁹⁶. Though Rituximab treatment was discontinued early in third trimester, after delivery one baby had about a 5% decrease in B cells, while the other baby had a complete loss of B cells⁹⁶. However, in both cases, babies were delivered at full term and required relatively minor follow-up care after delivery⁹⁶, suggesting the promise of these interventions individually or in combination. Because such case-studies are by nature anecdotal, future studies will be needed to rigorously establish the safety and benefit of these and other strategies to reduce HDFN.

Other approaches realizing success in allo- or auto-immune states that work by reducing levels or blocking the function of the pathogenic antibodies in circulation could also hold promise. To this end, Imlifidase, or IdeS, a selective IgG Fc domain protease repurposed from Streptococcus pyogenes has been used to desensitize organ transplant recipients who are positive for antibodies to donor organ antigens^105–108. If employed in the context of HDFN, this drug would be anticipated to both reduce maternal antibody levels as well as restrict placental transfer. Although care should be taken in trials of IdeS, as patients experienced unexpected severe complications in a small trial for treatment of immune thrombotic thrombocytopenic purpura (iTTP)¹⁰⁸. Another future prospect might be the selective clearance or blockade of pathogenic afucosylated antibodies, such as has been recently described in a mouse model of severe dengue infection using a nanobody specific for afucosylated IgG Fc¹⁰⁹. In sum, as compared to the paucity of data suggesting optimism for the prospects for prevention of sensitization, the pipeline of therapies that might reduce HDFN in the context of sensitized pregnancies clearly holds considerable promise.

6. Conclusion

The seriousness of K-mediated fetal anemia, the lack of prophylactic treatment to prevent K sensitization, and the limited tools available for treating K- mediated fetal anemia create a difficult situation for affected families. While a prophylactic polyclonal serum Ig preparation or mAb capable of preventing sensitization is desirable, the challenges to successful development and even deployment of such a treatment are significant. Despite decades of use, the mechanism by which RhIg protects remains unclear. Surrogate measures of reduced likelihood of sensitization have not been dependable⁵⁹, and translation from animal models to humans is not straightforward¹¹⁰. On the other hand, many characteristics, including epitope specificity, antibody subclass, and glycosylation state appear to affect mAb prophylactic action^66,73, leading to interesting new understanding of maternal and fetal immunology and antibody activity. Fortunately, new strategies for treating K-ADFN include blocking FcRn and reducing alloantibody in maternal circulation, which have the potential to treat multiple causes of antibody-mediated HDFN. Together these new developments point to a more hopeful future for affected families.

Highlights.

• Maternal sensitization to K may cause of severe anemic disease of the fetus and newborn.

• Candidate interventions used in research settings to prevent alloantigen sensitization sometimes have the opposite effect.

• Novel therapeutics to treat sensitized pregnancies aim to block maternal antibody transfer across the placenta.

• Reducing maternal IgG transfer in sensitized pregnancies may reduce pathology in K affected pregnancies.

List of abbreviations

α

anti

ADCC

antibody dependent cellular cytotoxicity

ADFN

anemic disease of the fetus and newborn

FcγR

Fcγ receptor

FcRn

neonatal Fc receptor

EGA

estimated gestational age

HDFN

hemolytic disease of the fetus and newborn

HOD

fusion protein of hen lysozyme, ovalbumin, and Duffy human antigen

IUT

intrauterine transfusion

IVIG

intravenous immunoglobulin

K

immunogenic Kell blood group antigen

mAb

monoclonal antibody

MZ

marginal zone

RBC

red blood cell

Rh

Rhesus antigen

D

Rhesus antigen D

RhIg

polyclonal α-Rh IgG, prophylactic treatment for D sensitization