Hematopoietic stem cell transplantation (HSCT) is a potentially curative treatment for a wide range of inherited and acquired immune and hematological diseases. Next to the aim to replace the affected hematopoietic system, a well-reconstituting immune system is pivotal for adequate long-term graft function and clinical outcome. Various factors, including match grade; graft source; conditioning regime; the occurrence of graft-versus-host-disease (GvHD); infections and chimerism shape the characteristics of immune reconstitution. While T-cell reconstitution has been studied extensively, studies on B-cell reconstitution are limited (1–3). Using a linear mixed model, our group has demonstrated that B-cell reconstitution is dependent on time post-transplant, donor age and recipient age – a host-factor –, but not on stem-cell source, match grade and conditioning regimen (4). Long-term immunological complications include selective hypogammaglobulinemia, autoimmunity and inadequate vaccination responses; all of which are associated with impaired functional B-cell recovery (5, 6).

The underlying mechanisms of hypogammaglobulinemia post-HSCT are not well understood. One important facet is the influence of cognate interactions between T and antigen-presenting cells affecting locally produced cytokines and chemokines that would normally drive terminal B-cell differentiation. Here, we describe a patient who suffered from hypogammaglobulinemia and mixed chimerism persisting for up to 7 years post-HSCT for sickle-cell disease (SCD) whom we have studied through extensive phenotyping of B-cell development in peripheral blood and secondary lymphoid organs to determine the nature of her hypogammaglobulinemia.

We report a 16-year-old female patient with SCD experiencing frequent vaso-occlusive bone crises who underwent HSCT from her mismatched sibling donor. Patient and donor shared the paternal, and part of the maternal human leucocyte antigen (HLA)-haplotype. A crossover event during meiosis-I caused the patient to receive the translocated HLA-DRB1 and HLA-DQB1 alleles from the other maternal haplotype.

The patient was preconditioned with fludarabine (cumulative dose 150mg/m²) and dexamethasone (cumulative dose 125mg/m²) and developed cervical lymphadenopathy later deemed fludarabine-associated toxicity after exclusion of an infectious cause. Subsequent conditioning included fludarabine, treosulfan and anti-thymocyte globulin (cumulative dose 150mg/m²; 42 gr/m²; and 8mg/kg, respectively).Cyclophosphamide was given at day 3-4 (cumulative dose 100mg/kg), followed by mycophenolate mofetil (30 mg/kg/day) with cyclosporin A (CsA; plasma levels 150-200 μg/L) from day +5 onwards as graft-versus-host-disease (GvHD) prophylaxis. Post-HSCT her lymphadenopathy recurred and considered as toxic or reactive to fludarabine after ultrasound scan. On day 19, CsA was switched to sirolimus due to inadequate plasma levels.

The patient’s cellular recovery was normal initially with normal T- and B-cell counts ( Figure 1A ) and without GvHD nor PCR-detectable viral reactivations. However, her humoral immunity did not appear to reconstitute, with IgG levels consistently dropping until below reference values at 3 months ensuing immunoglobulin replacement therapy (IGRT) which was started at 6 months post-HSCT. Following a period of patient-initiated reduced follow-up frequency, IGRT therapy was restarted with cotrimoxazole for frequent respiratory tract infections. It appeared her cervical lymphadenopathy never fully resolved, and she developed IgA-deficiency (0.15-0.16 g/L, ref 0.70-4.00 g/L) with a strong increase in IgM (7.29-8.11 g/L, ref 0.40-2.30 g/L) and was further diagnosed with acute cutaneous, discoid and chilblain lupus erythematosus with borderline positivity of anti-nuclear antibodies at 3 years post-HSCT. Donor peripheral blood chimerism decreased gradually from 99% at 2 months to 44% in granulocytes and 30% in leucocytes (37% in CD19⁺ B, 53% in CD4⁺ and 56% in CD8⁺ T cells) at 3-years post-transplant, remaining stable since. There were no SCD complaints since transplant at a stable HbS of 9%. No anti-cytokine antibodies against IL-6, IL-17a, IL-22, IFN-α and IFN-γ were detected. Only HLA class-I antibodies were found pre- and post-HSCT and the donor’s IgG, IgA and IgM titers were normal, so there was no clear explanation for the persistent hypogammaglobulinemia and clinical complaints from the patient’s diagnostic workup.

We describe a patient with persistent hypogammaglobulinemia, SLE-like skin lesions and lymphadenopathy after HSCT for SCD with mixed chimerism. We found no evidence of an intrinsic B-cell defect, because they expressed AID, could undergo SHM and were able to differentiate into IgG- and IgA-producing plasma cells in vitro. In the GC we found few, class-switched IgG+ plasma cells, indicating that terminal B-cell differentiation was not totally impaired but happened infrequently. Since all GCBCs were of patient origin, while only 26% of Tfh cells originated from patient, we hypothesize that the patient’s class-switch defect is related to HLA-II disparity limiting efficient T-B crosstalk between GCBCs and Tfh cells.

The patient had a translocation in the maternal HLA haplotype, receiving the HLA-II genes from the other maternal allele starting at HLA-DRB1. HLA class-I genes of patient and donor were fully matched, and at least half of the typed HLA-II alleles are matched (HLA-DRB1*14:54 and HLA-DQB1*05:03 of those tested). There is currently no evidence that HLA class-II disparity increases the risk for post-transplant hypogammaglobulinemia based on cohort studies on immunological recovery from mismatched related and unrelated grafts. In fact, mismatched related and unrelated donors were described to have similar B-cell recovery to matched related donors (14–16). Whether or not B-cell dysregulation occurs more frequently in patients with HLA class-II disparate HSCT and differential chimerism of T- and B-cell subsets remains to be determined.

Both patient and donor-derived naïve B cells were present in peripheral blood, indicating that the bone marrow microenvironment could support B-cell development of both patient and donor-derived B cells. Because we had no clinical indication to perform a bone marrow biopsy, we could not study the chimerism of B-cell precursors further. Analysis of bone marrow B-cell development can be especially informative in hypogammaglobulinemia cases coinciding with B-cell lymphopenia. Despite an ample supply of both patient and donor-derived naïve B cells in peripheral blood and LN, MBC-output was severely diminished. Naïve B cells were not intrinsically impaired in CSR in vitro, leading us to believe that terminal B-cell differentiation was impeded in the secondary lymphoid organs. In the LN, antigen-activated B cells migrate to the T-B border, where they activate CD4+ T cells. TCR signaling and ICOS-ICOSL crosslinking upregulates Bcl-6, CXCR5 and later IL-21, causing now pre-Tfh cells to mature and migrate into the GC (17). We could not deduce why Tfh cells were mainly donor-derived in the LN, but this must be due to an altered selection process.

Next to Tfh maturation, cognate TCR:pMHCII interaction also bifurcates B-cell differentiation into the extra-follicular (EF)- or GC-fate depending on the degree of T-cell help. A strong T-cell help signal is known to instigate GC formation and terminal differentiation of antigen-selected memory B cells, which depends on TCR:pMHCII crosslinking of multiple complexes (10, 18). The formation of GCs was not constrained by the noticeably skewed T:B chimerism, yet exclusively patient-derived B cells could initiate a GC reaction, even though 19% naïve donor B cells resided in the lymph node. The GCBC-fate, determined by intermediate T-cell help via CD40 and Bcl-6 was initiated by patient-derived B cells, but IRF4 expression was not sufficient to induce class-switching pre-GC maturation. Class-switching is dependent on strong T-cell help which facilitates AID-expression via IRF4 (19, 20), which we could normally simulate during B-cell differentiation in vitro. GC formation and architecture appeared normal in LN, and patient-derived GCBCs underwent normal proliferation and SHM compared to HDs in the DZ. In addition, antigen selection of the BCR repertoire was normal and apoptosis was not increased in the GC, indicating that cognate antigen on follicular dendritic cells (FDCs) provided GCBCs with the initial LZ-survival signal. Next, light-to-dark zone recirculation or differentiation into MBCs or plasma cells is again governed by T-cell help-levels.

Evidenced by sparse IRF4-expression, terminal plasma-cell differentiation was not blocked entirely but happened infrequently, indicating that patient GCBCs rarely received sufficient help from predominantly donor-derived Tfh cells (18, 21). While IgG⁺ but not IgA⁺ B cells were detectably by flow cytometry and immunohistochemistry, we found both IGHG and IGHA sequences by BCR repertoire sequencing. Based on the varying follicular hyperplasia of the lymph node tissue, it appears the remaining GCBCs become trapped in the GC, further increasing in size by the continual recruitment of additional activated B cells. While donor B cells did not initiate GCs they could form EF MBCs, a differentiation pathway that is not yet fully understood but likely results from weak T-cell help (18). Because MBCs in LN and peripheral blood had low SHMs and were mostly IGHM, we can conclude that patient and donor-derived B cells predominantly received weak T-cell help, via EF and/or GC responses.

HLA-compatibility plays a crucial role during the GC reaction, where high-affinity peptide-MHC-TCR interactions are necessary to induce terminal B-cell differentiation. HLA-II mismatches between centrocytes and FDCs may impede optimal antigen selection, although we would in principle expect 50% successful T:B interactions since at least the paternal HLA-II allotypes were matched between patient and donor. Evidently, haploinsufficiency alone might not be the sole explanation for the impediment in B-cell differentiation observed in this unique patient.

Based our investigations, we conclude that the patient’s hypogammaglobulinemia persisting for up to 7 years post-HSCT appears to result from inefficient T-B crosstalk inside the lymph node due to HLA-II disparity and skewed chimerism between T and B cells, although the exact mechanism remains elusive. We have shown that B-cell phenotyping of peripheral blood alone does not give the full picture of post-HSCT hypogammaglobulinemia; in depth cellular and molecular analysis of the GC reaction in secondary lymphoid organs can provide additional insights when studying these cases.

The IGH repertoire sequencing data have been deposited in the Gene Expression Omnibus (GEO) database and can be accessed through accession number GSE263200.

The studies involving humans were approved by the Institutional Review Board of the LUMC-WAKZ (B17.001). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

MG: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. IP-K: Investigation, Methodology, Writing – review & editing. MO-T: Investigation, Methodology, Writing – review & editing. DB: Writing – review & editing, Data curation. FS: Writing – review & editing, Data curation. AB: Writing – review & editing, Investigation, Methodology. HJ: Writing – review & editing, Data curation. PJ: Investigation, Methodology, Writing – review & editing. AL: Conceptualization, Supervision, Writing – review & editing. MB: Conceptualization, Methodology, Supervision, Writing – review & editing.