Partially Reversible Immunodeficiency After In-Utero Exposure to Immunosuppressants

Treatment with immunosuppressants affects pregnant women too. The European League Against Rheumatism (EULAR) has published recommendations for action for the use of immunosuppressants in pregnancy; azathioprine (n=72) and rituximab (n=256) were not associated with increased rates of miscarriage or malformations (1). But long term data on the effects on children’s immune reconstitution after exposure to these substances are lacking (2). We report on three cases of exposure to rituximab or azathioprine in utero that led to partly reversible, partly sustained immunodeficiency.

Methods

Collecting the clinical and immunological characteristics of three cases from the pediatric-immunologic outpatient clinics at Düsseldorf University Hospital and Hamburg-Eppendorf University Medical Center

Case descriptions and results

Patient 1: A pregnant 21 year old woman experienced a recurrence of bcr-abl positive c-ALL (acute lymphoblastic leukemia). Treatment with nine administrations of intravenous rituximab (375 mg/m² body surface) was initiated from the 22nd week of gestation. At birth (cesarean section in the 34 week owing to maternal indication) the otherwise health neonate had B-cell aplasia and lowered IgG (figure 1). After four months, B-cells and IgM rose to their age-appropriate standard values. In the absence of the patient’s own IgG production and a lacking response to vaccination, polyvalent immunoglobulins are still being substituted after five years. No class switched memory B-cells were confirmed by flow cytometry—that means only IgM producing memory B-cells were measurable, but no IgA, IgG, and IgE producing ones.

Figure 1.

Irreversible immunoglobulin class switch deficiency after exposure to rituximab in utero: immunoglobulin levels (y-axis) of subclasses IgG (dark red for patient 1, dark blue for patient 2) and IgM (orange for patient 1, light blue for patient 2) are plotted over the course of the patients’ ages (x-axis). IgA is not detectable in either patient at any point. The correlating duration of immunoglobulin substitution (grey bar), vaccinations (arrow, I = inactivated vaccine, L = live vaccine), and lacking confirmation of vaccine antibodies (star) are shown. Patient 1 and patient 2 were exposed to rituximab in utero.

Patient 2: A 34-year old pregnant woman was diagnosed with Burkitt lymphoma in the 19th week of gestation and she was subsequently treated with five cycles of R-CHOEP [rituximab CYCLOPHOSPHamide DOXOrubicin vinCRISTine etoposide prednisolone] (3). Her baby boy was born by cesarean section at 34 weeks of gestation because of a maternal indication and developed in an age-appropriate manner. At 16 months immunoglobulin substitution was initiated because he was susceptible to infections and had agammaglobulinemia (no confirmation of IgG, IgA, and vaccine antibodies, but normal IgM), which was continued after nine years when cessation attempts were unsuccessful (figure 1). Immunophenotyping showed normal B-cell counts but low memory B-cell counts.

Patient 3: A 34 year old woman who—eight years before her pregnancy—had received a kidney transplant because of terminal renal failure in a setting of renal malformation was treated with prednisolone, tacrolimus, and azathioprine during her pregnancy. Post partum, her healthy daughter who had been born at full maturity was found on neonatal pathology screening to have severe combined immunodeficiency (SCID). Immunologic diagnostic tests found severe B- and T-lymphopenia as well as undetectable IgM and IgA (figure 2). At 17 months her T-cell count had recovered only but she had standard B-cell counts and B-cell differentiation. Immunoglobulin substitution was required, and over time IgA, IgM, and IgG values rose spontaneously in the setting of an adequate response to vaccination. Initial pharmacologic diagnostic tests confirmed the azathioprine metabolite 6-thioguanin nucleotide in the neonatal blood. The TPMT*1/3*A-genotype or TPMT*3B/3C-genotype was confirmed in the child, which is associated with a slow metabolism of thiopurine methyltransferase (TPMT) and which is most likely to have caused the transplacental cell-mediated lymphotoxicity. In all three children, genetic (whole exome based) diagnostic testing for congenital immunodeficiencies yielded normal results.

Figure 2.

T-cell immune reconstitution after exposure to azathioprine in utero: the total count of all T-cells (CD3+, grey), T-helper-cells (CD3+ CD4+, dark blue), and cytotoxic T-cells (CD3+ CD8+, dark red) (y-axis) are plotted in months over course of the patients’ age (x-axis). Patient 3 was exposed to azathioprine during the pregnancy.

Discussion

Prenatal maternal immunosuppressive therapy can lead in the neonate to the immunologic phenotype of a primary immunodeficiency. An irreversible B-cell class switch deficiency (lacking synthesis of alternative antibody isotypes, such as IgG, IgA, and IgE) after exposure to rituximab in utero has been described in only one case report to date (4), whereas in numerous studies the B-cell depletion was reversible after six months. The anti-CD20+ directed IgG 1-antibody rituximab is known to cross the placental barrier and in the fetus increases to 120% of the maternal concentration (40th week of gestation) (4). From the 12th week of gestation the fetus produces CD20+ positive pre-B-cells and for the second and third trimester a vulnerable phase is therefore postulated during which a high risk exists for developing an irreversible class switch deficiency. Whether further genetic factors facilitate this development could not be determined in the present cases. Prophylactic immunoglobulin substitution is required as the babies—if left untreated—would have been at great risk for severe respiratory infections. In 2019 neonatal screening for SCID was introduced nationwide in Germany. PCR based measurements of the T-cell receptor-excision circles—a surrogate for the production capacity of naive T-cells—identify neonates with severe T-cell deficiency. Only few of the examined children were found with abnormalities on screening after exposure to immunosuppressant drugs in utero. Timing, duration, dosage, and the type of immunosuppressant probably affect the surrogate marker to differing degrees. Secondary T-cell deficiency in neonates is not the target of the screening. Without prophylactic measures, the outcome may be life threatening infections. In the third case, treatment of the mother with azathioprine caused—because of a genetic predisposition for a slow metabolizer of azathioprine—a partially reversible phenotype of a severe T-cell and B-cell deficiency in utero, which was detected thanks to neonatal screening. Azathioprine crosses the placental barrier; the fetal blood concentration is 5% of the maternal level. Case reports of this constellation have been published in the Netherlands (5). Because of these three cases we consider routine follow-up care of neonates exposed to immunosuppressants in a pediatric immunology department a sensible step. Especially before live vaccines are administered, the immunophenotypic cell characterization should be complemented by simultaneously undertaking functional assays. Before immunosuppressants are given during pregnancy, the parents should be comprehensively informed about possible positive TREC screening. Genetic testing of both parents and possibly also the child for the TPMT genotype by means of maternal blood should also be considered.