Successful pregnancy in the setting of eosinophil depletion by benralizumab

With widespread availability of mAbs for the treatment of eosinophilic disorders affecting women of childbearing age, safety during pregnancy has become an important issue. Although the package inserts for mepolizumab, reslizumab, and benralizumab indicate insufficient safety data in pregnancy, these drugs have demonstrated great benefit in the treatment of severe asthma and are likely to be prescribed to women of childbearing age. We describe a patient with hypereosinophilic syndrome who became pregnant while receiving benralizumab on a clinical trial and delivered a healthy baby without eosinophils. Nonhuman primate data supporting the safety of benralizumab during pregnancy are also provided.

The mother is a 36-year-old woman with hypereosinophilic syndrome and severe eosinophilic gastrointestinal involvement (see Kuang et al¹ for her clinical history and benralizumab response before becoming pregnant [patient 16]). She was unexpectedly found to be pregnant at a routine benralizumab administration visit (week 128; NCT02130882). After discussion with the National Institutes of Health Institutional Review Board, the National Institutes of Health Investigational New Drug sponsor, and AstraZeneca, she was allowed to continue benralizumab therapy. At 38 weeks of gestation (July 18, 2019), she delivered a 3091-g healthy baby girl by primary cesarean section for failure to progress. APGAR scores were 8 and 9 at 1 and 5 minutes, respectively. Physical examination of the neonate was entirely normal. Complete blood cell count at birth showed a white blood cell count of 22,500/μL with 0 eosinophils/μL. State-mandated newborn screening revealed a mildly elevated thyroid-stimulating hormone level (with normal free T4 level) that decreased over the ensuing months. The patient elected not to breast-feed. The baby’s growth and development have been within normal limits. She has been healthy (no sick visits) without eczema, food allergy, or other evidence of atopic disease. Her absolute eosinophil count remained undetectable until age 7 months at which time it was 228/μL (normal range for age, 700–1000/μL; Figure 1). Repeat absolute eosinophil count at age 1 year was 258/μL. Platelets began to rise at age 1 month, peaking at age 7 months at 723 k/μL before returning to near-normal values (492 k/μL) at 1 year. The remainder of her complete blood cell count has been consistently within normal limits (see Table E1 in this article’s Online Repository at www.jaci-inpractice.org). Peripheral blood smear review and flow cytometric staining of whole blood at age 1 year confirmed the presence of phenotypically normal eosinophils. She received pneumococcal conjugate vaccine at 4, 6, 9, and 12 months. Protective pneumococcal antibody levels to 12 of 12 tested serotypes included in the pneumococcal conjugate vaccine 13 and 0 of 11 tested serotypes not included in the vaccine were demonstrated at 12 months (see Table E2 in this article’s Online Repository at www.jaci-inpractice.org), consistent with an appropriate vaccine response.

FIGURE 1.

Transient suppression of eosinophilia in a baby born to a mother receiving benralizumab. Eosinophil count, hemoglobin (open circles), and platelet count (gray circles) over time are shown. The horizontal dashed line represents the lower level of normal for eosinophil count between birth and 18 months at the hospital where the testing was performed. AEC, Absolute eosinophil count; HgB, hemoglobin; Plat, platelet.

TABLE E1.

Laboratory parameters over time for human infant

Month	WBC (k/μL)	Hgb (g/dL)	MCV (fL)	Platelet (k/μL)	ANC (k/μL)	ALC (k/μL)	AMC (k/μL)	AEC (k/μL)	ABC (k/μL)	TSH (mIU/L)	Free T4 (ng/dL)
0	22.5	16.9	ND	336	15.52	4.73	1.575	0	0	35.6 (<35.0)	17.4 (>5)
1	9.74 (5.00–19.50)	10.3 (10.0–18.0)	98.7 (85.0–123.0)	429 (150–400)	1.85 (1.00–9.00)	6.92 (2.50–16.50)	0.97 (0.20–2.40)	0 (0.70–1.00)	0 (0.00–0.20)	5.12 (0.47–4.68)	1.15 (0.65–1.85)
4	8.2 (5.00–19.50)	11.3 (9.5–13.5)	85.8 (74.0–102.0)	521 (150–400)	1.26 (1.00–9.00)	5.66 (2.50–16.50)	1.27 (0.20–2.40)	0 (0.70–1.00)	0.01 (0.00–0.20)	6.81 (0.47–4.68)	1.06 (0.65–1.85)
7	13.98 (6.00–17.50)	11.8 (10.5–13.5)	85.6 (70.0–86.0)	723 (150–400)	3.22 (1.00–8.50)	10.07 (4.00–13.50)	0.42 (0.20–2.40)	0.28 (0.70–1.00)	0 (0.00–0.20)
12	12.91 (6.00–17.50)	11.8 (10.5–13.5)	85.3 (70.0–86.0)	492 (150–400)	2.40 (1.00–8.50)	8.40 (1.00–10.50)	0.96 (0.10–1.10)	0.24 (0.70–1.00)	0.13 (0.00–0.20
13.5	12.76 (6.00–17.50)	12.4 (10.5–13.5)	85.7 (70.0–86.0)	436 (150–400)	2.81 (1.00–8.50)	9.31 (1.00–10.50)	0.38 (0.10–1.10)	0.26 (0.70–1.00)	0 (0.00–0.20

ABC, absolute basophil count; AEC, absolute eosinophil count; ALC, absolute lymphocyte count; AMC, absolute monocyte count; ANC, absolute neutrophil count; CBC, complete blood cell count; HgB, hemoglobin; MCV, mean corpuscular volume; ND, not done/determined; TSH, thyroid-stimulating hormone; WBC, white blood cell.

Normal ranges are given in parentheses below each value.

Abnormal values are indicated in bold type.

Blood draw for initial TSH and T4 testing was performed at 24.5 h of life. Initial testing was performed at the Arkansas Department of Health. CBCs and subsequent thyroid testing were performed at Arkansas Children’s Northwest Laboratories.

TABLE E2.

Pneumococcal titers at age 12 mo for human infant

Pneumococcal serotype*	Prevaccination IgG (μg/mL)	Postvaccination IgG (μg/mL)	Fold rise in antibody concentration to PCV13 serotypes
1	4.9	6.98	1.4
2	0.11	0.15
3	1.64	3.14	1.9
4	4.6	16.09	3.5
5	5.76	13.60	2.4
6A	Not tested	Not tested
6B	2.36	13.39	5.7
7F	1.0	1.72	1.7
8	0.17	0.23
9N	0.93	2.79
9V	1.91	5.39	2.8
10A	0.5	0.7
11A	0.18	0.17
12F	0.06	0.05
14	1.16	4.56	3.9
15B	0.05	0.07
17F	0.68	0.73
18C	5.09	13.51	2.7
19A	1.93	3.26	1.7
19F	4.4	9.43	2.1
20	0.78	1.01
22F	0.32	0.3
23F	3.9	20.98	5.4
33F	0.06	0.07

PCV13, pneumococcal conjugate vaccine 13.

Serotypes contained in pneumococcal conjugate vaccine (PCV) 13 are indicated in bold. Assay was performed at ARUP Laboratories.

^*A response to 50%–70% or more of the serotypes in the vaccine, defined as an antibody concentration >1.0 μg/mL, is considered protective. A 2- to 4-fold rise in antibody concentration postvaccination is considered a response.

A Good Laboratory Practice–compliant enhanced prenatal and postnatal development study was performed in cynomolgus monkeys. Pregnant monkeys received benralizumab 10 mg/kg (n = 14), 30 mg/kg (n = 21), or control (n = 15) every 14 days beginning at day 20 to 22 of gestation through delivery (for additional details, see this article’s Online Repository text at www.jaci-inpractice.org). Growth and development were normal throughout gestation and postnatal follow-up (6.5 months) in the monkeys exposed to benralizumab in utero. The only benralizumab-related effect was depletion of eosinophils, which resolved by 180 days of life in all but 1 infant monkey (in the 30 mg/kg exposure group) (Figure 2). No abnormalities were observed in other leukocyte subsets, hemoglobin levels, or platelets counts. Benralizumab administration resulted in depletion of bone marrow eosinophils in the same infant monkey that lacked detectable peripheral blood eosinophils. No other bone marrow changes were observed. Serum levels of IgG, IgM, and IgA and the T-cell–dependent antibody response to primary keyhole limpet hemocyanin immunization were within normal limits and comparable to control animals. Serum half-life of repeated dose intravenous benralizumab (10–30 mg/kg) in adult cynomolgus monkeys is 11 to 17 days (compared with ∼15 days in adult humans²).

FIGURE 2.

Blood eosinophil counts in cynomolgus monkeys born to mothers receiving benralizumab (10 mg/kg or 30 mg/kg) or control vehicle. Individual counts over time are shown in gray with dashed lines. Geometric mean values are shown in black with solid lines. AEC, Absolute eosinophil count; Benra, benralizumab.

Although murine models have demonstrated no evidence of teratogenicity or reproductive toxicity in the absence of eosinophils, eosinophil-less mice have been shown to have subtle abnormalities of homeostatic function, including impaired recall responses to vaccines.³ These findings raise theoretical concerns for babies born to mothers receiving eosinophil-depleting antibodies that cross the placenta, including benralizumab. Since Food and Drug Administration approval for use in severe asthma in adults and children 12 years or older in November 2017, more than 8000 individuals have received benralizumab in clinical trials. To date, there have been no reports of unexpected toxicities, and a large study examining antibody responses to seasonal influenza vaccination showed no impairment of responses in patients with asthma receiving benralizumab.⁴ The only published report of benralizumab in human pregnancy to date is an abstract describing a woman with marked eosinophilia (6940/μL), asthma, multifocal pulmonary infiltrates, and eosinophilic pericarditis, well controlled on benralizumab, who relapsed with drug discontinuation.⁵ Benralizumab was restarted at 20 weeks of gestation and continued through parturition. No further information is provided.

Although limited to a single case, the current report clearly demonstrates that eosinophils are not essential for normal human development in utero through age 1 year. The 7-month delay in detectable peripheral eosinophils is consistent with the nonhuman primate postpartum data but slightly longer than the 3 to 6 months reported in benralizumab-treated patients following discontinuation of therapy.^6–8 Elevated eosinophil counts at age 3 months have been associated with atopic disease in childhood.⁹ It is interesting that, despite a strong family history of early atopic disease, the human child has had no evidence of atopy to date. Although the transient increased platelet count remains unexplained, reactive thrombocytosis is common in the first 2 years of life and typically resolves spontaneously without clinical consequences. Thrombocytosis was not observed in the benralizumab-exposed infant monkeys.

In summary, we report the case of a healthy child born to a mother receiving benralizumab throughout pregnancy and provide data from preclinical pregnancy studies in nonhuman primates supporting the hypothesis that eosinophils are not essential for normal human development. Although no safety signals were identified, more data are clearly needed to establish the risks and benefits of benralizumab and other eosinophil-depleting therapies in pregnancy.

ONLINE REPOSITORY

ASSESSMENT OF EOSINOPHILS

Peripheral blood smears were stained with Wright-Giemsa for morphology evaluation using standard procedures. Multiparameter flow cytometry of the peripheral blood was performed using directly conjugated antibodies against CD9, CD11b, CD13, CD34, CD69, and HLA-DR (Beckman Coulter, Miami, Fla), CD16, CD45, CD117, and IL-5Rα (Becton Dickinson, San Jose, Calif), sialic acid–binding Ig-like lectin 8 (BioLegend, San Diego, Calif), and EMR-1 (Bio-Rad, Hercules, Calif). Acquisition was performed using the standard 10-parameter, 8-fluorescent channel FACS Canto II cytometer (Becton Dickinson). Flow cytometry data were analyzed using FCS Express 6 Analysis Software (Glendale, Calif). Eosinophils were defined as cells positive for sialic acid–binding Ig-like lectin 8, EMR-1, CD9, IL-5Rα, and CD45 and negative for CD34, CD117, and CD16.

BENRALIZUMAB MATERNAL, EMBRYOFETAL, AND NEONATAL TOXICITY STUDY

The cynomolgus monkey was selected for nonclinical safety evaluation on the basis of binding of benralizumab to cynomolgus monkey eosinophils, the pharmacological activity of benralizumab in an IL-5–induced model of eosinophilia in cynomolgus monkeys, and the demonstrated antiasthmatic activity of benralizumab in a cynomolgus monkey asthma model.^E1 This study was performed with female cynomolgus monkeys at an Association for Assessment and Accreditation of Laboratory Animal Care–accredited laboratory.

Parameters evaluated for adult females included clinical signs, changes in food consumption, body weight, pregnancy status, and clinical pathology. Embryofetal parameters included measurements by ultrasound and gestational length. The pregnant females were allowed to deliver their infants by natural birth. The day of parturition was considered birth day (BD) 1 for the infants. Parameters evaluated for neonates (infants) through BD28 included clinical signs; changes in body weight; infant physical, external, morphometric, muscle tone, neurobehavioral, ophthalmologic, and skeletal examinations/assessments; and changes in peripheral blood lymphocytes, and immunoglobulin levels.

After BD28, for an additional 5.5 months postpartum/postnatal, the adult females and infants continued to be evaluated for clinical signs and changes in food consumption (adults), body weight, hematology parameters, peripheral blood lymphocytes, and immunoglobulin levels. Infants continued to be evaluated for external, morphometric, and ophthalmologic examinations/assessments, in addition to immunologic assessment of T-cell– dependent antibody response to keyhole limpet hemocyanin (KLH) (5 mg KLH was administered to each infant by intramuscular injection on BD180 ± 1 day. Blood samples were taken pre-KLH challenge and 7, 10, and 14 days after KLH challenge for measurement of KLH IgM and IgG. On BD199 (±2 days), infants underwent a visceral examination and full necropsy, including macroscopic tissue examinations. A subset of tissues was weighed, and tissues were collected, preserved, and microscopically examined.

Clinical Implications.

• Eosinophil-depleting therapies have shown benefit in women of childbearing age. Although the risks and benefits of these agents in pregnant women with hypereosinophilia have not been established, this study demonstrates that eosinophils are not required for normal human development.