Neonatal Alloimmune Thrombocytopenia With Amegakaryocytosis, B Lymphopenia, and Villitis

A healthy-looking term infant, a girl who was the third child of a healthy and nonconsanguineous couple, was admitted to the neonatal department for isolated profound thrombocytopenia (11 G/L), which was revealed by a complete blood cell (CBC) count performed at birth. The first child was a healthy boy. Past medical familial history was remarkable for the death of the second child (newborn 2), who was a girl born after an uncomplicated pregnancy, after 12 hours of life because of a massive cerebral hemorrhage. The CBC count revealed deep thrombocytopenia (6 G/L). The autopsy showed amegakaryocytosis on the medulla in the biopsy, villitis on the placenta (Fig 1), lymphocytic depletion both on the spleen and lymph nodes, and hemophagocytosis on the medulla, lymph nodes, and liver samples. The diagnosis of neonatal alloimmune thrombocytopenia (NAIT) was not further investigated because fetal-maternal antigen incompatibility was found in the class I human platelet antigen (HPA) system (mother HPA-1a/a, newborn HPA-1a/b) and was not associated with circulating or fixed antibodies, in neither the newborn nor the mother. There was no incompatibility in HPA-3 and HPA-5 systems. Maternal serum immunologic testing (including antinuclear antibody, extractable nuclear antigen, anti–double stranded DNA, anti–thyroid peroxidase, antithyroglobulin, and antiphospholipid) were all negative. The mother’s platelet count was normal, eliminating autoimmune thrombocytopenia. Infection screens (syphilis, toxoplasmosis, rubella, cytomegalovirus, and parvovirus B19) were also negative. The patient reported here (newborn 3) was born 2 years later; a CBC count at birth revealed deep thrombocytopenia (11 G/L). Clinically, only a few petechial hemorrhages were noted.

Given the sibling history, further explorations were undertaken; bone marrow smears revealed isolated amegakaryocytosis without hemophagocytosis. Lymphocytes analysis showed complete B lymphocytopenia (CD19⁺ B-lymphocytes = 0, confirmed on a second sample few days later), moderate T- and natural killer (NK)-lymphocytopenia (CD3⁺ T lymphocytes = 1200/µL [normal 2500–5500], CD16⁺ CD56⁺ NK lymphocytes = 78/µL [normal 170–1100]), and an absence of T-lymphocyte activation (CD3⁺DR⁺ lymphocytes <1% [normal <1%], CD4⁺DR⁺ lymphocytes <1% [normal <1%], and CD8⁺DR⁺ lymphocytes <1% [normal <1%]). Additionally, both perforine expression on NK cells and T-lymphocyte degranulation were found normal. Strikingly, as for newborn 2, microscopic examination showed major lymphoid infiltration involving the villous placenta. Again, screening tests for NAIT investigating HPA-1, HPA-3, and HPA-5 systems were negative. Treatment consisted in an intravenous immunoglobulin perfusion (1 g/kg) and an alternate donor platelet transfusion (day 1). This resulted in significant but transient elevation in the platelet count (80 G/L). Two additional random donor platelet transfusions were given at day 6 and day 14 to maintain a platelet count over 20 G/L. No hemorrhagic complication occurred.

In the presence of thrombocytopenia, hemophagocytosis signs observed on newborn 2’s tissues and lymphoid depletion in newborns 2 and 3 first raised the hypothesis of hemophagocytic lymphohistiocytosis (HLH).¹ Moreover, the early neonatal presentation and familial recurrence were indicative of familial HLH. However, impairment of lymphocyte cytolytic activity, the hallmark of all genetic disorders that are associated with familial HLH, was absent in our patient. In addition, hemophagocytosis was seen exclusively in the first affected neonate, whereas thrombocytopenia was profound in both neonates, which also suggested that hemophagocytosis was not the cause of thrombocytopenia. Finally, screening for mutations in the genes known to be associated with familial HLH was not considered. Amegakaryocytosis was consistent with congenital disorder of glycosylation type 1a, but gene sequencing was found normal. Although none of the inherited thrombocytopenias were known to include B lymphocytopenia, gene sequencing for CMPL, AML1, and WASP expression studies were performed, but results, months later, proved to be negative.

The normalization of both the B-lymphocyte count (at 15 days of life) and the platelet count (at 2 months of life) was indicative of a transient disorder and prompted us to consider NAIT in an unexplored system. New samples obtained after personal communication with a referenced testing laboratory finally allowed the diagnosis of NAIT in a HPA-15 system by revealing a fetal-maternal alloantigen incompatibility (mother: HPA-15a/a; father: HPA-15a/b; newborn 3: HPA-15a/b) and detecting in maternal serum circulating antiplatelet antibodies. These antibodies were shown to be specific to HPA-15. An identical HPA-15a/b genotype was retrospectively found in newborn 2, consistent with the diagnosis of NAIT in both siblings.

Villitis of unknown etiology (VUE) is clearly distinct from diffuse sclerosing villitis because of infectious agents such as rubella or toxoplasmosis. It is characterized by a major destructive inflammatory lesion involving the villous placenta. In addition to CD8+ maternal T cells, resident placental macrophages of fetal origin are key participants in the inflammatory process.² There are increasing arguments for a maternal immune response against the foreign fetal allograft.^3,4 Because of the well-known association between VUE and maternal autoimmune diseases,⁵ the mother had been, after the death of the second child, thoroughly tested for underlying autoimmune disease. The link between VUE and alloimmunization is more recently but now clearly emerging.^6,7 We acknowledge that the detection of VUE for the first affected child should have led us to investigate alloimmune disorders and furthered the hypothesis of NAIT.

The presentation of NAIT was atypical in our case because thrombocytopenia was associated with amegakaryocytosis and B lymphocytopenia. The basis for the association between NAIT and B lymphocytopenia (newborn 3) and lymphocytic depletion in the spleen and lymph nodes (newborn 2) is speculative. It is possible that an increased expression of HLA antigen class II molecules, which allows an enhanced potential for antigen presentation by fetal macrophages as shown in recent studies elucidating VUE,⁴ could result in allograft rejection by the mother against fetal B lymphocytes expressing HLA antigen class II antigens. Unfortunately, in our cases, blood sampling was no longer available, precluding this hypothesis from being tested. Alternatively, B lymphopenia may be the result of an aberration in the cytokine modulation. We thus believe that functional tests for the identification of maternal antibodies against HLA antigen class II molecules may provide a clue to the pathophysiology of B lymphopenia and should be performed in addition to detection of HLA antigen class I alloantibodies.

Our study also uncovers another novel feature of NAIT, revealing amegakaryocytosis in both siblings. Thrombocytopenia in NAIT is supposed to be mediated by accelerated clearance of antibody-opsonized fetal platelets. Our observations indicate that thrombocytopenia could result from a nonexclusive peripheral mechanism and that megakaryocytes could be a direct target of HPA-15 antibodies. Interestingly, recent studies describe maternal anti–HPA-1 antibodies that can suppress fetal megakaryopoiesis.⁸ Whether a similar effect can be attributed to anti–HPA-15 antibodies remains to be investigated. In addition to alloimmunization against platelet antigens, neonatal thrombocytopenia might result from platelet consumption linked to endothelial cell damage within the villous capillaries.⁶ Abnormal thrombogenesis has been observed in severe VUE and is associated with intrauterine growth restriction.⁷ In our report, the siblings were not growth retarded and no vaso-occlusive lesions were identified in both placentas.