Neonatal-Onset Refractory Thrombocytopenia and Anemia Associated With a Homozygous In-Frame NFE2 Duplication

Introduction

NFE2 encodes p45/NF-E2, a terminal megakaryocyte transcription factor required for proplatelet formation, a conclusion strongly supported by animal genetics. However, human congenital cytopenia attributable to NFE2 variation remains exceptionally rare, and the current clinical signal is still defined largely by isolated reports – predominantly loss-of-function – that establish the initial disease association [1–6]. Accordingly, domain-adjacent, non-truncating variation in NFE2, particularly when presented at birth, represents an important interpretive gap. In this context, we report the case of a neonate with day-1-onset bicytopenia, marrow dysmegakaryopoiesis with reduced late megakaryocyte forms, and non-response to thrombopoietin-receptor agonism, in whom trio testing identified a recessively segregating, bZIP-proximal in-frame NFE2 duplication. This clinicogenetic scenario broadens both the genotypic spectrum (a short in-frame duplication adjacent to a critical domain) and the phenotypic spectrum (neonatal bicytopenia), while providing a testable mechanism of disrupted bZIP-dependent transcriptional activation.

More broadly, inherited bone marrow failure syndromes (IBMFS) are a clinically and genetically heterogeneous group of disorders defined by lineage-specific or multilineage cytopenias, variably accompanied by congenital anomalies and an increased lifetime risk of myelodysplastic syndrome (MDS), acute myeloid leukemia (AML), and solid malignancies. Critically, IBMFS can present in the neonatal period without overt syndromic stigmata; consequently, refractory cytopenias can be misattributed to “acquired” neonatal etiologies, delaying definitive counseling and mechanism-aligned care [7–10]. At the molecular level, IBMFS converge on a limited number of core pathways – including DNA repair deficiency (Fanconi/BRCA), telomere biology disorders (eg, DKC1, TERT, TPP1/ACD), and ribosome biogenesis defects (eg, DBA/SDS) – with TP53-linked stress responses contributing to ineffective hematopoiesis [8,11,12] Telomere-pathway disorders have age-spanning expressivity, ranging from ultra-rare neonatal neurodevelopmental phenotypes to adult aplastic anemia [12–14] Because phenotypes and natural histories overlap, early germline evaluation that pairs next-generation sequencing (NGS) approaches with parallel copy-number variant (CNV) analysis has become central to accurate diagnosis, variant reinterpretation, and genetic counseling [15–17] Moreover, precise subtyping informs curative decision-making; hematopoietic stem cell transplantation (HSCT) remains definitive for many IBMFS, with outcomes influenced by donor/source selection and conditioning strategy [18,19]

Within this framework, platelet biogenesis provides a mechanistically tractable bridge between genotype and phenotype. Hematopoietic stem cell-derived megakaryocyte–erythroid progenitors (MEPs) commit to megakaryopoiesis under thrombopoietin (THPO)–MPL signaling, undergo endomitosis and cytoplasmic maturation, and ultimately elaborate proplatelets that release circulating platelets; transcription factors including GATA-1, FLI1, and p45/NF-E2 coordinate this late maturation program [1,2] Model systems reinforce conservation of thrombopoiesis and clarify pathway-level vulnerabilities, including thpo-deficient zebrafish with marked thrombocytopenia [20] and murine transcription factor module perturbations (SCL/Lyl1 with reduced Nfe2 expression) that compromise proplatelet formation [21,22]. Epigenetic interference (HDAC inhibition) further highlights the sensitivity of the GATA-1/NF-E2 axis during terminal megakaryocyte maturation [23]. Animal genetics place NF-E2 at the terminal step of megakaryopoiesis: NF-E2–deficient mice lack circulating platelets and die of hemorrhage despite preserved responsiveness to THPO, supporting a late maturation/proplatelet defect rather than an upstream cytokine axis failure [3,24]. Beyond platelet quantity, NF-E2 programs platelet-function genes (eg, SELP, Myl9) and require obligate small Maf partnerships; additionally, SUMOylation of p45 at K368 enhances DNA binding and transactivation [5,6]. Human evidence remains limited; nonetheless, a homozygous NF-E2 loss-of-function variant causing severe early-infancy thrombocytopenia with megakaryocytic maturation defects has established NF-E2 as a bona fide congenital thrombocytopenia gene [25]. Importantly, NF-E2 biology plausibly extends beyond platelet biogenesis because NF-E2 and GATA-1 coordinate shared terminal maturation programs across megakaryocytic and erythroid lineages, and NF-E2-null models demonstrate direct effects on erythropoiesis and globin gene regulation [1,5,26,27].

Clinically, refractory neonatal cytopenias are frequently approached initially as immune-mediated or consumptive processes; consequently, delayed recognition of germline etiologies can prolong ineffective therapy, defer mechanism-directed surveillance, and complicate anticipatory counseling and family planning. This report therefore emphasizes a practical diagnostic pivot point – persistent cytopenias accompanied by dysmegakaryopoiesis and thrombopoietin-receptor agonist refractoriness – in which early trio testing with CNV analysis may be decisive for establishing etiology and guiding management [27]. Accordingly, we present this case to expand the emerging NF-E2–associated disease spectrum to neonatal bicytopenia, to underscore the pathogenic plausibility of a short bZIP-proximal in-frame duplication, and to highlight the immediate clinical value of comprehensive germline testing in neonatal bone marrow failure contexts [7,10,27].

Case Report

CLINICAL PRESENTATION AND INITIAL EVALUATION:

A late-preterm female infant born to consanguineous parents developed day-1–onset bicytopenia, dominated by severe thrombocytopenia with concomitant anemia, while leukocyte counts remained preserved. Dysmorphic features were noted (short, webbed neck); however, chromosomal analysis confirmed 46,XX, and early growth and feeding were appropriate. Initial laboratory testing demonstrated platelets of approximately 20×109/L with hemoglobin near 10 g/dL. Coagulation indices and metabolic screening were unremarkable, and cranial ultrasonography showed no intracranial hemorrhage. Because an alloimmune mechanism was initially favored, intravenous immunoglobulin was administered for presumed neonatal alloimmune thrombocytopenia; nevertheless, no hematologic response was observed, and platelet transfusions produced only transient increases, with progressive transfusion dependence (Figure 1).

HEMATOLOGIC INVESTIGATIONS:

Given the persistence of cytopenias despite immune-directed therapy, a bone marrow examination was performed. Marrow cellularity was appropriate for age, with active granulopoiesis and erythropoiesis; however, megakaryopoiesis was abnormal, showing dysmegakaryopoiesis with reduced late megakaryocyte forms, alongside a maturation lag in erythroid precursors. Blasts were 2% by morphology and <1% CD34+ by flow cytometry, and an MDS panel was negative. Collectively, these findings shifted the working diagnosis away from consumptive or alloimmune thrombocytopenia and toward an inherited defect affecting terminal megakaryocyte maturation.

GENOMIC INVESTIGATIONS AND SEGREGATION:

Trio-based exome sequencing, supplemented by proband whole-genome sequencing, identified 3 clinically relevant variants. First, a homozygous PKHD1 pathogenic variant (c.4870C>T; p.Arg1624Trp) accounted for the hepatorenal phenotype, consistent with ARPKD/Caroli disease. Second, a homozygous in-frame 12-bp duplication in NFE2 (c.889_900dup; p.Glu297_Arg300dup) was detected and classified as a variant of uncertain significance (VUS) (Figure 2). To contextualize this finding within p45/NF-E2 domain architecture, we mapped the duplication relative to the CNC and bZIP modules (Figure 3A, 3B). Third, a heterozygous ABCB4 variant (c.1529A>G; p.Asn510Ser) was interpreted as likely pathogenic and consistent with heterozygous carrier status in the absence of supportive phenotypic correlation. Parental testing confirmed heterozygosity for the PKHD1 and NFE2 variants in both parents, supporting autosomal-recessive segregation in this consanguineous family. The father was negative for the ABCB4 variant, whereas the mother was heterozygous.

From a hematologic perspective, the NFE2 duplication localizes to the C-terminal portion of p45/NF-E2 within the CNC region (aa 287–329) and adjacent to the basic DNA-binding region (aa 330–343) and leucine zipper (aa 344–373) (Figure 3A, 3B). IGV review demonstrated a read pattern consistent with a biallelic 12-bp duplication with defined breakpoints (Figure 2), and copy-number analysis was negative. The duplication is not reported in ClinVar and is classified as a VUS in VarSome and Franklin. Variant classification followed ACMG/AMP criteria with ClinGen Sequence Variant Interpretation refinements. The homozygous in-frame NFE2 duplication is absent or ultra-rare in population databases; accordingly, we applied PM2_Supporting. The duplication localizes within the CNC region adjacent to the basic leucine zipper module that is central to NF-E2 DNA binding and small Maf partner dimerization; therefore, we applied PM1_Supporting (domain relevance without an established human mutational hotspot). Segregation is consistent with autosomal-recessive inheritance; however, because the evidence is derived from a limited number of informative meiosis in a single nuclear family, we applied PP1 at a conservative supporting strength. Computational evidence supports a deleterious effect (PP3_Supporting), and the phenotype – neonatal-onset persistent thrombocytopenia with dysmegakaryopoiesis and lack of sustained response to TPO-receptor agonist therapy – provides PP4_Supporting. Functional evidence is currently unavailable; consequently, PS3/BS3 are not applied, and there are insufficient independent cases to support PS4. Taken together, the available evidence remains insufficient for (likely) pathogenic classification; thus, the variant is retained as a Variant of Uncertain Significance (VUS) pending functional confirmation.

MANAGEMENT RATIONALE AND CLINICAL COURSE:

Clinically, the infant required recurrent platelet and red-blood-cell transfusions, with only brief post-transfusion platelet rises. Because marrow findings suggested impaired megakaryocyte maturation rather than cytokine deficiency, a therapeutic trial of romiplostim (a thrombopoietin-receptor agonist, TPO-RA) was initiated as a time-limited mechanistic trial to assess cytokine axis responsiveness; however, there was no sustained platelet response, and transfusion dependence persisted. Follow-up is reported from January 2022 through January 2025 (approximately 36 months), during which thrombocytopenia remained persistent with only transient post-transfusion increases and no sustained response to romiplostim (Figure 1); the sequence of major diagnostic and therapeutic events is summarized in Figure 4. Darbepoetin alfa (an erythropoiesis-stimulating agent, ESA) was used to support erythropoiesis and reduce red-cell exposure, and deferasirox was introduced to mitigate transfusional iron accumulation, with serial ferritin monitoring and routine renal/hepatic toxicity surveillance. In parallel, the confirmed PKHD1-related ARPKD/Caroli disease prompted coordinated nephrology and hepatology surveillance, including blood-pressure management, renal function monitoring, and hepatobiliary follow-up. The ABCB4 heterozygous finding required no direct intervention beyond counseling.

DIFFERENTIAL DIAGNOSIS AND MECHANISTIC INTERPRETATION:

In neonatal practice, immune-mediated and consumptive causes of thrombocytopenia are appropriately prioritized. Accordingly, neonatal alloimmune thrombocytopenia (NAIT) was considered given the day-1-onset severe thrombocytopenia; maternal anti-human platelet antigen (HPA) antibody testing was performed, and the patient was treated empirically with intravenous immunoglobulin (IVIG) while results were pending. Subsequently, the NAIT evaluation returned negative, and there was no meaningful platelet count response to IVIG; therefore, an alloimmune mechanism became less likely and the diagnostic focus shifted toward inherited etiologies. The key diagnostic and therapeutic milestones are summarized in Figure 4. In this setting, persistent bicytopenia accompanied by dysmegakaryopoiesis should prompt early escalation to a germline cytopenia framework. Moreover, the combination of day-1-onset cytopenias, marrow evidence of impaired terminal megakaryocyte maturation, and refractoriness to romiplostim (a thrombopoietin-receptor agonist, TPO-RA) argues against a primary THPO deficiency state or other isolated upstream cytokine insufficiency. Rather, these findings support a defect downstream of THPO-MPL signaling, in which cytokine pathway augmentation cannot restore platelet production because the terminal megakaryocyte program is not effectively executed. Notably, NFE2 encodes p45/NF-E2, a terminal megakaryocyte transcription factor required for proplatelet formation; therefore, the recessively segregating NFE2 in-frame duplication proximal to the bZIP module provides a coherent and testable hypothesis for impaired transcriptional activation with consequent failure of late megakaryopoiesis (Figure 3A, 3B).

GENETIC COUNSELING AND ETHICS:

Given the recency and breadth of whole-genome sequencing, repeat testing was not pursued in the absence of a major phenotype shift. The family was counseled regarding the incomplete diagnostic yield of current genomic strategies for congenital cytopenias and the practical implications of a VUS, which cannot be used as a decision-directing marker. A prior pregnancy had undergone testing and was terminated due to PKHD1 homozygosity; therefore, counseling emphasized the feasibility of PGT-M/IVF for PKHD1 and discussed the limitations of incorporating the NFE2 VUS beyond research and longitudinal reinterpretation. Written informed consent for genetic testing and publication was obtained, and all data were handled in accordance with institutional ethics standards.

Discussion

Congenital disease attributable to NF-E2 (encoded by NFE2) is very rare. To date, only a single infant with homozygous loss-of-function has been described, presenting with profound early thrombocytopenia and defective megakaryocytic maturation [25]. Our patient differs in 2 important respects. First, the clinical picture was not isolated thrombocytopenia but neonatal-onset bicytopenia with intermittent anemia. Second, the genetic lesion is not truncating but a short homozygous in-frame duplication located within the CNC region and immediately adjacent to the basic DNA-binding and leucine zipper elements. Taken together, these observations suggest that the disease spectrum associated with NFE2 disruption may be broader than previously appreciated and that perturbation of the C-terminal regulatory architecture – even without truncation – may be biologically consequential. This interpretation is consistent with longstanding experimental data placing NF-E2 at the terminal step of megakaryocyte differentiation [3,27,28] and is aligned with observations from contemporary inherited bone marrow failure syndrome (IBMFS) cohorts, in which early-onset cytopenias frequently reflect genetically heterogeneous etiologies [16].

The observed duplication permits a mechanistic interpretation rather than a purely descriptive association. The NFE2 in-frame duplication (p.Glu297_Arg300dup) is positioned immediately proximal to the bZIP module of p45/NF-E2, the structural interface responsible for DNA binding and heterodimerization with small Maf partners (MafG/MafK), which are essential for NF-E2 transcriptional activity [28,29]. This C-terminal region is structurally constrained; even short in-frame insertions may alter local spacing across the CNC–basic–LZ architecture and potentially affect heptad alignment or partner interaction. Such perturbations could impair transcriptional activation of genes required for late megakaryocyte maturation and proplatelet formation [3,5,27,29,30].

Experimental models provide contextual support for this interpretation. Murine studies demonstrate that NF-E2 deficiency results in terminal megakaryocytic arrest with failure of proplatelet formation despite preserved thrombopoietin responsiveness, favoring a downstream maturation block rather than a primary cytokine axis defect [24,26,27]. In addition to regulating platelet number, NF-E2 transactivates genes central to platelet function, including SELP and MYL9, linking its activity to both quantitative and qualitative platelet output [5]. The functional sensitivity of this region is further supported by evidence that C-terminal integrity is modulated by post-translational mechanisms, such as SUMOylation near the C-terminus, suggesting that non-truncating alterations in proximity to the bZIP domain have biological consequences [5,6]. The intermittent anemia and erythroid maturation lag observed in our patient are also biologically plausible in this framework, as NF-E2 contributes to erythroid gene regulation and shared terminal differentiation programs across megakaryocytic and erythroid lineages in experimental systems [3,4].

From a clinical perspective, persistent neonatal thrombocytopenia – particularly when accompanied by anemia and dysmegakaryopoiesis – should prompt early reconsideration of germline etiologies. Immune-mediated and consumptive causes are appropriately prioritized in neonatal practice, yet failure to achieve sustained increments despite transfusion, together with marrow findings demonstrating impaired late megakaryocyte maturation, represents a practical inflection point toward an inherited bone marrow failure syndrome framework [7–9,17]. In such settings, early trio-based germline sequencing with parallel CNV analysis can be decisive, not simply for diagnosis but also for counseling and therapeutic planning [7–9,16–18].

The lack of response to thrombopoietin-receptor agonist therapy in this case is clinically instructive. THPO-MPL signaling maintains megakaryocyte mass and platelet production when the downstream differentiation program is intact [1,2]. Experimental work demonstrates that ligand-level deficiencies can respond to restoration or mimicry of cytokine signaling [20]. In contrast, absence of response despite adequate stimulation raises concern for impairment beyond the cytokine axis. This pattern aligns more closely with a terminal transcriptional block at the level of NF-E2 rather than an upstream signaling deficiency [24,26,27]. In neonates who fail to respond to TPO-RA therapy and demonstrate marrow dysmegakaryopoiesis, consideration should therefore extend beyond THPO deficiency to include MPL-associated disorders, MECOM-regulated states, and transcriptional regulators of terminal megakaryopoiesis [17,18,31–34].

We have intentionally retained a Variant of Uncertain Significance classification and made the evidentiary framework explicit. The duplication appears to be very rare or absent in population datasets, supporting PM2 [35]. Its localization within a functionally critical C-terminal module supports PM1 at supporting-to-moderate strength [5,24,26,27,29,30]. Autosomal-recessive segregation in the nuclear family supports PP1 (moderate), and the phenotypic specificity supports PP4, in agreement with some previous studies [4,7,8]. However, PS3 cannot be invoked in the absence of functional data, and PS4 is not applicable without independent enrichment. Functional studies remain essential. A rigorous approach would include transactivation assays, DNA-binding assessments of canonical response elements, evaluation of small Maf partner interaction, and downstream transcriptional profiling under controlled expression conditions [5,29,32,35]. Only concordant functional impairment across orthogonal systems would justify upgrading this allele beyond VUS.

Segregation consistent with autosomal-recessive inheritance carries a 25% recurrence risk and a 50% carrier probability for siblings [7,8]. These data support targeted family counseling and discussion of reproductive options, while acknowledging that VUS status constrains decision-directing application pending further evidence [7,8,16,17]. Broad germline testing with concurrent CNV analysis has become increasingly central in neonatal cytopenias, as structural variants represent a meaningful diagnostic fraction [16,17,36].

Management remains focused on bleeding mitigation, rational transfusion strategy, monitoring for alloimmunization and iron overload, and vigilance for infectious complications, with consideration of hematopoietic stem cell transplantation if progressive marrow failure emerges [4,11,18,19,27]. Standard IBMFS surveillance principles apply [7,8,18].

The principal limitation of this report is the absence of functional validation in a single patient. While the mechanistic plausibility is strong, causal inference remains provisional. Independent case identification and experimental confirmation will be necessary to clarify the pathogenic role of bZIP-proximal in-frame alterations in NFE2.

Conclusions

We describe the case of a neonate with persistent bicytopenia, dysmegakaryopoiesis, and a homozygous in-frame NFE2 duplication positioned adjacent to the bZIP module. The clinical phenotype and refractoriness to thrombopoietin-receptor agonism support impaired terminal megakaryocytic maturation rather than an upstream cytokine axis deficiency. Although the variant remains a Variant of Uncertain Significance pending functional validation, its positional and biological plausibility justify targeted functional studies and further case aggregation. Clinically, this case reinforces the value of early trio-based germline testing with concurrent CNV analysis in refractory neonatal cytopenias.