LTβR deficiency causes lymph node aplasia and impaired B-cell differentiation

Abstract

Secondary lymphoid organs (SLOs) provide the confined microenvironment required for stromal cells to interact with immune cells to initiate adaptive immune responses resulting in B-cell differentiation. Here, we studied three patients from two families with functional hyposplenism, absence of tonsils and complete lymph node aplasia, leading to recurrent bacterial and viral infections. We identified biallelic loss-of-function mutations in LTBR encoding the lymphotoxin beta receptor (LTβR), primarily expressed on stromal cells. LTβR-deficient patients had hypogammaglobulinemia, diminished memory B cells, regulatory and follicular T-helper cells, and dysregulated expression of several Tumor Necrosis Factor family members. B-cell differentiation in an ex vivo co-culture system was intact, implying that the observed B-cell defects were not intrinsic in nature, and instead resulted from LTβR-dependent stromal cell interaction signaling critical for SLO formation. Collectively, we define a human inborn error of immunity caused primarily by a stromal defect affecting the development and function of SLOs.

Introduction

Secondary lymphoid organs (SLOs) are strategically localized throughout the body in the form of tonsils, lymph nodes, the spleen, and Peyer’s patches (1). They act as surveillance centres and provide the specialized microenvironment necessary for the initiation of adaptive immune responses (2). Through a complex network of stromal cells alongside tightly regulated chemical signals, they facilitate the interaction of various immune cell types, culminating in the formation of germinal centers (GCs) where high-affinity antibody secreting plasma cells and memory B cells are formed (3).

Patients with inborn errors of immunity (IEIs) such as BTK (4), RAG1 or RAG2 (5) deficiencies can present with non-palpable lymph nodes. However, in these cases, the stromal compartment develops normally, and the lymph node structures are formed, although they are not populated by lymphocytes. Following successful hematopoietic stem cell transplantation (HSCT), these lymph nodes can organize and function properly (6).

Still, there are rare cases of IEIs with aberrant SLO development beyond defects in lymphocytic compartments. Patients deficient in NF-κB-inducing kinase (NIK) exhibit lymph node aplasia (7), while 40S ribosomal protein SA (RPSA) deficiency is characterized by isolated congenital asplenia (8). Notably, NIK is ubiquitously expressed and hence directly influences lymphocyte function and development, whereas RPSA deficiency does not affect lymph node architecture or other lymphoid organs beyond the spleen (7, 8). As far as we are aware, no isolated defect causing IEI by disrupting the stromal architecture of SLOs has been described (9).

Nuclear factor-κB (NF-κB) is a family of transcription factors with critical roles ranging from the coordination of immune and inflammatory responses in both innate and adaptive immune systems to the development and maintenance of lymphoid organs (10). Several receptors can activate NF-κB signaling in immune cells, including the T-cell receptor (TCR), B-cell receptor (BCR), Toll-like receptors (TLRs) and members of the tumor necrosis factor (TNF) superfamily (10). The lymphotoxin beta receptor (LTβR), a member of the TNF superfamily, is primarily expressed by stromal cells such as endothelial, mesenchymal, and epithelial cells, as well as by myeloid cells including dendritic cells (DC) and macrophages (11, 12). In contrast, its two ligands, lymphotoxin (LTα1β2) and TNFSF14/LIGHT, are predominantly expressed on activated lymphocytes and lymphoid tissue inducer cells (11, 13). Notably, these inducer cells have been demonstrated to be required for lymph node formation during embryonic development in mice (14). Upon ligand binding, LTβR activates both the canonical NF-κB pathway and, to a greater extent, the non-canonical NF-κB pathway (15). The pivotal role of LTβR in the development and regulation of the immune system, through tight communication between stromal cells and lymphocytes, has been established in murine models where its absence (16) or inhibition (17) results in absent lymph nodes and Peyer’s patches as well as an impaired splenic architecture. However, the role of LTβR signaling in humans has remained poorly defined.

Here, we report an IEI caused by biallelic loss-of-function (LOF) mutations affecting LTβR, and reveal the distinct role of LTβR signaling governing stromal-cell function in shaping the microarchitecture and immunological functions of SLOs in humans previously unappreciated in murine studies.

Results

An inborn error of immunity with lymph node and tonsil aplasia and splenic defect

We studied three male patients (P1-P3) from two unrelated consanguineous families (Fig. 1A) who had recurrent upper and lower respiratory tract infections starting 4-6 months after birth, predominantly of bacterial etiology, and requiring intravenous antibiotic treatment. At the age of 3 years, P1 experienced meningitis caused by S. pneumoniae and recovered completely. P2 had an episode of acute hepatitis at the age of 9 years. No causative agent was identified, and a liver biopsy revealed biliary destruction (Fig. S1A). We detected an accumulation of CD4⁺ T cells, B cells, and some CD8⁺ T cells in the patient’s biopsy, reminiscent of lymphoid infiltrates observed in studies of Ltbr^−/− mice (16, 18). Following antibiotic treatment and cholecystectomy, the patient no longer exhibited symptoms of acute hepatitis. The older brother of P3 had similar disease manifestations and succumbed to disease complications, including pulmonary hypertension and cor pulmonale, at 18 years of age. Despite the recurrent infections, lymphadenopathy was not detected in P1-P3, and physical examination was remarkable for absent tonsils and non-palpable lymph nodes. Clinical histories are summarized in Table 1 and supplemental patient clinical histories.

Figure 1. Identification of LTβR-deficient patients.

(A) Pedigrees of the two unrelated families included in this study. Black solid symbols indicate affected individuals. Genotypes are indicated below the symbols. Squares indicate male and circles female family members. Slashed symbols indicate that the individual has died. Roman numerals indicate generations, and Arabic numbers indicate individuals within a generation. (B) Lymphoscintigraphy images depicting the lower lymphatic system in a healthy control (HC) alongside P1 and P2 (posterior view). P1 and P2 display normal lymphatic duct development but lack inguinal and iliac lymph nodes (blue dashed rectangle). Blue arrows indicate the main lymphatic nodes. The red arrow indicates the injection sites. (C) Howell-Jolly bodies (blue arrows) in erythrocytes in a blood smear from P2. (D) Geometric Mean Fluorescence Intensity (gMFI) of CD47 in lymphocytes. Data representative of N=2; HC (n=7) and patients (n=3). Statistical analysis performed on one of these experiments using Unpaired t-test with Welch’s correction. (E) Innumerous verrucae planae (flat warts) on the neck of P1. (F) Serum values for CXCL13 for P1-P2 and controls (n=4) from two separate Luminex multiplex assays (complete results displayed in Fig. S4). (G) LTβR expression by flow cytometry analysis in fibroblasts from a HC, P1-P3, as well as in P1-derived fibroblasts where the mutation was reverted to wild type by CRISPR-Cas9 editing (CRISPR-KI). (H) Representative immunoblot displaying the expression of p52 before and after stimulation with the lymphotoxin (LT) ligand in a HC, P1 and P1-derived CRISPR-KI fibroblasts. Quantification shown as fold change of p52/HSP90 relative to untreated HC1. Heat shock protein 90 (HSP90) serves as loading control. Unpaired t-test was used for statistical analysis in panel D and F, and G.

Table 1. Clinical and immunological features of LTβR deficient patients.

N.d. Not detected; n.a. Not available; * on IgRT; N.d. Not detected; n.a. Not available; TREC T-cell receptor excision circle; KREC kappa-deleting recombination excision circle; TREC and KREC reference values from (86), WBC whole blood cell count; ALC absolute lymphocyte count; ANC absolute neutrophil count; ATC absolute thrombocyte count, ENA extractable nuclear antigen, LSC Light Signal Count

	Patient 1			Patient 2				Patient 3
LTBR genomic change (HGVS, NC_000012.12)	g.6384449C>T			g.6384449C>T				g.6385266G>C
LTBR CDS change (HGVS, ENST00000228918.9)	c.91C>T			c.91C>T				c.359G>C
LTBR protein change (HGVS, ENSP00000228918.4)	p.Gln31Ter			p.Gln31Ter				p.Arg120Pro
Age (years), Sex	26, Male			19, Male				9, Male
Age (months) at onset	7			3				6
Age at evaluation (years)	11	17	25	4	8	13	17	6 mo	1.5	8
WBC (cells x10^3/mm3) Normal Range	6.1 4.5-13.5	4.3 4.5-13	7.0 4.5-11	5.7 5-15.5	5.7 4.5-13.5	4.7 4.5-13.5	5.8 4.5-13	22.6 6-17.5	12.0 6-17.5	9.8 4.5-13.5
ALC (cells x10^3/mm3) Normal Range	2.1 1.5-6.5	2.1 1.2-5.2	2.0 1-4.8	4.0 2-8	4.0 1.5-6.8	2.8 1.5-6.5	2.6 1.2-5.2	15.0 4-13.5	11.0 4-10.5	2.5 1.5-6.8
ANC (cells x10^3/mm3) Normal Range	3.6 1.5-8.5	1.6 1.8-8	4.1 1.8-7.7	1.4 1.5-8.5	1.4 1.5-8	1.5 1.5-8.5	2.7 1.8-8
ATC (cells x10^3/mm3)) Normal Range	229 150-350	218 150-350	195 150-350	295 150-350	315 150-350	490 150-350	297 150-350
IgA (mg/dl) Normal Range	<25 (67-433)	10 (139-378)	<27 (139-378)	24 (57-282)	60 (78-383)	4 (96-465)	27 (139-378)	<6.67 (7-123)	<6.67 (30-307)	<6.67 (78-383)
IgG (mg/dl) Normal Range	690 (835-2694)	1700* (913-1884)	1060* (913-1884)	558 (745-1804)	851 (764-2134)	370* (987-1958)	1180* (913-1884)	150 (304-1231)	630* (605-1430)	1000* (764-2134)
IgM (mg/dl) Normal Range	27 (47-484)	31 (88-322)	37 (88-322)	64 (78-261)	<17 (69-387)	NA	38 (88-322)	45 (32-263)	50 (66-228)	60 (69-387)
IgE (lU/mL) Normal Range	12 <60	13.4 <60	89 <60	35 <60	5 <60	NA	9.72 <60	<5 <15	<5 <60	<5 <60
Anti-nuclear Antibody	n.a.	n.a.	Negative	n.a.	n.a.	n.a.	Negative	n.a.	n.a.	AMA-M2, Anti-Ku
ENA Profile	n.a.	n.a.	Negative	n.a.	n.a.	n.a.	Negative	n.a.	n.a.	n.a.
Anti Thyroid Antibody	n.a.	n.a.	Negative	n.a.	n.a.	n.a.	Negative	n.a.	n.a.	Negative
Isohemagglutinin Normal Range	1/64 >1/8	1/64 >1/8	1/8 >1/8	½ >1/8	n.a.	n.a.	½ >1/8	n.a.	n.a.	n.a.
TREC Normal Range (86)	912 Median (min-max): 6620 (1160-19 600)			572 Median (min-max): 12 600 (1120-36 200)				n.a.
KREC Normal Range (86)	725 Median (min-max): 1940 (1660-15 800)			520 Median (min-max): 11 335 (1720-61 000)				n.a.
Anti-IFN-α antibody Normal range < 1980LSC	n.a.			n.a.				121
Anti-IFN-ω antibody Normal range < 1961 LSC	n.a.			n.a.				110
Anti-IFN-γ antibody Normal range < 1516 LSC	n.a.			n.a.				78

Lymphoscintigraphy in P1-P3 showed complete lymph node aplasia despite normal lymphatic duct development (Fig. 1B, and Fig. S1B). The spleen had normal dimensions and morphology on ultrasonographic examination, but Howell-Jolly bodies in peripheral blood smears indicated severe functional hyposplenism (Fig. 1C, and Fig. S1C) (19). Additionally, expression of CD47, a signal used by cells to protect them from splenic removal by macrophages and dendritic cells (20) was decreased on lymphocytes from P1-P3, further highlighting impaired splenic function observed in LTβR-deficient patients (Fig. 1D).

Laboratory studies revealed low levels of IgA and IgG in P1-P3 (Table 1). IgM levels were below the age-adjusted normal range in P1 and P2, but normal in P3. Following a clinical diagnosis of combined immunodeficiency, regular intravenous immunoglobulin (IVIG) substitution and antibiotic prophylaxis was initiated in P1-P3, reducing exacerbations of respiratory tract infections. P1 tested DNA-positive for genus beta human papillomavirus type 24 (HPV-24) (Fig. 1E), which is associated with epidermodysplasia verruciformis (EV) and skin cancer (21). No other persistent viral infections were observed (Table 1, and supplemental patient clinical histories). Collectively, these results reveal an error of immunity with complete lymph node and tonsils aplasia and splenic defect.

Biallelic germline LTBR mutations resulting in loss of expression and impaired non-canonical NF-κB signaling

Given the enigmatic disease etiology, we performed whole-exome sequencing (WES) of P2 and P3 (Fig. S2, and Table S1-S4) and uncovered rare homozygous variants in the Lymphotoxin beta receptor (LTBR) gene, segregating with the disease under the assumption of autosomal recessive inheritance (Fig. 1A, and Fig. S3A). The LTBR variant in P1 and P2 introduces a premature stop codon in exon 1 (c.91C>T, p.Gln31Ter) and P3 carries a missense variant in exon 4 (c.359G>C, p.Arg120Pro), located in a loop of the second cysteine rich domain (Fig. S3B) and predicted to destabilize the protein due to steric clashes of the substituting proline residue (Fig. S3C). The identified variants were either absent or ultra-rare in public databases (allele frequency < 0.00001), consistent with LTBR constraint metrics hinting at intolerance to both missense and LOF variation. In silico predictions suggested a likely deleterious effect of these variants (Table S5).

Serum analysis of P1 and P2 showed a decrease of B lymphocyte chemoattractant (CXCL13/BLC), a chemokine that is highly expressed in SLOs, where it controls the navigation of B cells (Fig. 1F, and Fig. S4A-C) (22, 23). CXCL13 is secreted following LTβR activation (24) and its reduction has been used as a serum biomarker of successful LTβR inhibition in clinical trials of LTβR antagonists (25, 26). Dermal fibroblasts of P1-P3 showed complete absence of LTβR protein expression (Fig. 1G). Upon binding its ligand lymphotoxin LTα1LTβ2 complex (LT), LTβR activates the non-canonical NF-κB pathway during which the precursor p100 is processed into the active p52 form (15). Accordingly, treatment of patient-derived fibroblasts with LT failed to upregulate p52 (Fig. 1H, and Fig. S5A). In contrast, canonical NF-κB pathway activation induced by the Tumor Necrosis Factor alpha (TNF-α) remained intact (Fig. S5B). Correction of the stop-gain variant in fibroblasts of P1 and P3 using CRISPR/Cas9 restored both LTβR expression and LT-induced p100 processing into p52 (Fig. 1G-H, Fig. S5A and Fig. S5C). Thus, these results demonstrate that the variants were LOF and causative of the observed, aberrant non-canonical NF-κB signaling.

Deficiency of memory B cells, regulatory and follicular helper T cells

Serial laboratory analyses for P1-P3 revealed normal ranges of total leukocyte and lymphocyte counts (Table 1). Despite LTβR being absent in B and T lymphocytes (11, 12, 27), as terminal B cell maturation occurs mainly in SLOs, we hypothesized that B-cell differentiation may be impaired (3). Correspondingly, despite normal total numbers of CD19⁺ B cells, we detected a significant reduction of GC-like B cells (Fig. 2A), as well as a near absence of both class-switched and unswitched memory B cells (Fig. 2B, and Fig. S6) and IgA⁺ or IgG⁺ B cells (Fig. S7A). Expansion of T-bet^highCD21^low B cells is a hallmark of chronic activation of the adaptive immune system in certain infections or autoimmune disorders and these numbers can be aberrant in patients with IEI (28). However, both P2 and P3 exhibited similar numbers of T-bet^highCD21^low B cells compared to healthy controls (HCs) (Fig. S7B). Following in vitro cytokine stimulation, patient B cells showed normal activation, proliferation, and in principle, the ability to undergo class-switch into IgA or IgG positive cells (Fig. S8). To assess the contribution of T cells to B-cell dysfunction, we analyzed the T-cell compartment and function in LTβR deficient patients. Mouse studies demonstrated the role of LTβR signaling in regulating thymic epithelial cells and stromal cells (18, 29). Although P1 and P2 showed lower T-cell Receptor Excision Circle (TREC) levels, their total number of T cells were in the normal range (Table 1), and P1-3 had comparable proportions of recent thymic emigrants to healthy controls (Fig. S9A). While the CD4⁺ T-cell subpopulations were unaffected, P1 and P2 displayed increased CD8⁺ terminally differentiated effector memory T cells re-expressing CD45RA (TEMRA) (Fig. 2C), which may correlate with chronic antigenic exposure (30). T-regulatory cells (Tregs) and T-follicular helper cells (TfH) were significantly reduced in P1-P3 (Fig. 2D, and Fig. 2E). Murine studies have, thus far, displayed conflicting results regarding the impact of lymphotoxin (LT) on T helper (Th) cell differentiation (13). While LT has been identified as a prototypical cytokine associated with Th1 cell responses (31), observations in mice deficient in LTβR or its ligand have showed elevated levels of Th1-type cytokines within their spleens and lungs (32). Conversely, exposure to Leishmania major infection resulted in a propensity towards Th2 cell polarization in Ltbr^−/− mice, resulting in an increase in the severity of the systemic infection (33). Stimulation of lymphocytes and assessment of cytokine production revealed a significant reduction of IL-4 producing Th2 cells in P1-3, as well as a trend towards lower IFN-γ-producing Th1 and IL-17A producing Th17 cells (Fig. 2F). P1 and P2 T lymphocytes exhibited similar functionality to those from HCs across various in vitro functional assays, including assessments of proliferation and activation, and naïve T-cell differentiation into Tregs (Fig. S9B, and Fig. S10).

Figure 2. In-depth characterization of the immune cell compartment from LT βR-deficient patients.

(A) Gating strategy to identify subpopulations of CD19⁺ B cells using flow cytometry and results from patients and HCs (n=9). Subpopulations were classified as Plasmablasts (PB), Pre-/ and Germinal-center (Pre-GC, and GC) Naïve or Memory/DN (double negative). Data shown here is from one experiment, representative of N=5; HC (n=7) and patients (n=3). (B) Representative plots from two HCs and P1-P3, highlighting the reduction of both class-switched (IgD^-CD27⁺) and unswitched (IgD⁺CD27⁺) CD19⁺ B cells in the patients. (C) Distribution of naïve (CD45RA⁺CCR7⁺), central memory (TCM CD45RA^-CCR7⁺), terminally differentiated effector memory T cells re-expressing CD45RA (TEMRA CD45RA⁺CCR7^-) and effector memory (TEM CD45RA^-CCR7^-) subpopulations within CD4⁺ and CD8⁺ T-cell fractions. The data shown here is from one experiment, representative of N=4; HC (n=7) and patients (n=3). (D) Frequency of CD25⁺FOXP3⁺ T-regulatory (Treg) cells. The data shown here is from one experiment, representative of N=2; HC (n=7) and patients (n=3). Statistical analysis was performed on one of these experiments. (E) Frequency of CD45RA^-CCR7^-CXCR5⁺ T-follicular helper (TfH) cells. The data shown here is from one experiment, representative of N=5; HC (n=7) and patients (n=3). Statistical analysis was performed on one of these experiments. (F) Distribution of T-helper (Th) subsets, characterized by cytokine expression on CD3⁺CD4⁺CD25^-CD45RA^- T cells following 5-hour stimulation with phorbol myristate acetate (PMA) and ionomycin. Results show a shift in the patients from IL-4 producing Th2 and IL-17 producing Th17 subsets towards the IFN-γ producing Th1 cells (HCs n=10). The experiment was performed twice with cells from P1 and P2. Statistical analysis was done using analysis of variance (ANOVA) followed by Bonferroni correction in (A) and Unpaired t-test with Welch’s correction in (D-F).

Single-cell RNA sequencing (scRNAseq) of lymphocytes of P1 and P2 confirmed the shift in the CD8⁺ compartment towards effector memory cells and a lower proportion of Tregs (Fig. 3A, Fig. 3B and Fig. S11). Among the most differentially downregulated genes in CD8⁺ T lymphocytes were LTB, encoding for lymphotoxin-β (LT-β) which binds to lymphotoxin-α (LT-α) to form the main ligand for LTβR, the lymphotoxin heterotrimer LTα1β2 or LTα2β1, and FOS, which plays a major role in response to antigenic activation (Fig. 3C) (34, 35). Despite a history of recurrent infections in both patients, differential gene expression analysis revealed no overt differences in the expression of genes associated with chronic inflammation and exhaustion pathways (Table S6). Clonality via bulk TCR sequencing (Fig. 3D) and scRNAseq analysis (Fig. S12) showed a reduction in the diversity of T cells from P1 and P2, however this was driven by the expansion of a few clones on a polyclonal background. Together, these findings confirm a phenotypic shift in T cells towards an effector memory population with a constrained clonotype arrangement, suggesting their relatively quiescent state despite repetitive antigenic challenge. The conserved in vitro functions of patient lymphocytes in contrast to altered in vivo differentiation imply that the observed alterations in subpopulation distributions may not stem from an intrinsic defect within lymphocytes themselves, but rather from alterations in the stromal compartment and the SLOs of the patients.

Figure 3. Sequencing analysis reveals LTB downregulation and reduced TCR clonality.

(A) Low-dimensional projection (uniform manifold approximation and projection (UMAP) plot) of the combined scRNAseq data set from P1 and P2 and four HCs. In the left panel, colors correspond to the cell type identified by label transfer from a reference data set of healthy PBMCs. The right panel displays the same projection, with colors indicating the distribution of patient cells (blue) with the clusters compared to controls (grey). (B) Distribution of CD8 T-cell compartments in the scRNAseq data. (C) Heatmap showing pseudobulk expression data for genes (shown as columns) that are differentially expressed between P1, P2 and HCs within the CD8+ T-cell population (samples and cell types shown as rows). Numbers on the left indicate cell count per group. Dot plots indicate the degree of significance, and in which subtypes they were observed. (D) TCR clonality results obtained from bulk DNA TCRB sequencing in the 1000 most abundant clones with a productive TCRB VDJ rearrangement and Shannon’s evenness indices below each plot.

Alterations in the LTβR and TNF network

Having noticed the significant downregulation of LTB expression across multiple CD8⁺ cell subtypes from P1 and P2 in the scRNAseq (Fig. 3C), we measured LTB serum levels in P2 and P3 which were significantly lower compared to healthy controls (Fig. 4B). We then assessed other members of the TNF family as well. Serum analysis showed a modest increase of TNF-α in P1 and P2 (Fig. 4A, and Fig. S13A), as well as TNF-β (the soluble homotrimer of LT-α) in P1 (Fig. 4A). TNF superfamily member 14 (TNFSF14/LIGHT), the other known ligand of LTβR, was also elevated, which was further confirmed by ELISA in all three patients (Fig. 4C) (12). Additionally, Fas Ligand (FasL) was also increased in the patients (Fig. 4D). Similar to TNF-α and TNF-β/LTα3 (36), both LIGHT(37) and FasL(38) can also induce apoptosis. Furthermore, overexpression of LIGHT in mice results in autoimmunity (39) and increased levels of LIGHT or FasL are associated with autoimmunity in humans (40, 41). In previous studies, Ltbr^−/− mice exhibited a phenotype of immune dysregulation characterized by splenomegaly, autoantibody production, and lymphocytic infiltrates (16, 18, 42). None of the patients in our cohort had clinically overt autoimmunity or autoinflammation. We screened the patients with extensive autoantibody detection panels, which were negative in P1 and P2. However, P3 tested positive for anti-Ku and anti-mitochondrial antibodies (AMA-M2) despite normal serum liver and renal parameters (Table 1) implying that the presence of these autoantibodies has not resulted in any clinical manifestations or organ-related disease in this patient to date. While autoantibodies against type I IFNs have recently been reported in several monogenic IEIs affecting the non-canonical NF-κB pathway (43), we could not detect these antibodies in serum from P3 (Table 1).

Figure 4. Alterations in the LTβR/TNF network and the immunomodulatory effect of DcR3.

(A) Heatmap depicting LEGENDplex™ multiplex assay for serum samples from P1 and P2 compared to HC (n=2). Results were individually normalized, with the lowest value set to 0 and the highest value set to 100. All other values were scaled proportionally between these two extremes. Numbers shown are absolute values in pg/mL. (B) ELISA serum analysis for lymphotoxin beta (LTB). Data from one experiment with HCs (n=7) and from P2 (n=2) and P3 (n=1). (C-E) ELISA results for LIGHT/TNFSF14 (C), FasL (D) and DcR3 (E) for serum samples from HCs (n=6) and samples pooled from P1 and P2, from two blood draws and one from P3. In (E) commercially available intravenous immunoglobulin (IVIG) was also tested. Samples from P1 and P2 were taken from two timepoints. (F) ELISA results of DcR3 fold change in treated fibroblasts normalized to untreated cells for HC (n=2), patients (n=3), and cells from P2 where TNFR1 was knocked out. Each datapoint represents the average of N=2. (G) Activation-induced cell death (AICD) following restimulation of feeder-expanded T cells with soluble anti-CD3 (sCD3). Graph representing the percentage of apoptotic cells with or without DcR3 treatment. Dots represent average of individual healthy donors (n=3) over two independent experiments. (H) Effect of DcR3 treatment on the killing capacity of expanded T-cells, based on % of surviving co-cultured p815 target cells treated with 0.1 sCD3 compared to the results in the same condition being treated with 0 sCD3. Data representative of two independent experiments, including two of the HCs repeated as a biological replicate. Statistical analysis performed on one of these experiments. Analysis in (B) was performed using Unpaired t-test with Welch’s correction. Analysis in (C-E) and (G) was performed using Unpaired t-test. Analysis in (F) and (H) was performed using ANOVA followed by Dunnett’s post-hoc test for multiple comparisons.

Additionally, increased levels of Decoy Receptor 3 (DcR3/TNFRSF6B) were observed in the patient’s serum compared to HCs (Fig. 4E, and Fig. S13B). DcR3 is a soluble protein with anti-inflammatory properties, acting through the inhibition of both LIGHT and FASL (44, 45). Surprisingly, we were also able to detect DcR3 in various samples of commercially available IVIG products (Fig. 4D, Fig. S13B, and Fig. S13C). Furthermore, we observed that IVIG treatment of an IEI patient resulted in increased DcR3 serum levels (Fig. S13C). Previous studies have linked DcR3 production to PI3K/NF-κB activation (46). Treatment of dermal fibroblasts with either TNF-α or TNF-β/LTα3 resulted in the secretion of DcR3 (Fig. 4F) as did the two known isoforms of lymphotoxin, LTα1β2 and LTα2β1. This effect was dependent on TNFR1 and LTβR stimulation respectively, as fibroblasts from P2 with an additional TNFR1 knockout showed no response to any of the ligands. It is important to note that while LTα2β1 is capable of binding both receptors, LTα1β2 binds exclusively to LTβR (47). Subsequently, LTα1β2 had no effect on patient derived fibroblasts, whereas LTα2β1 was also able to induce DcR3 production to a comparably to that of healthy controls. To assess the immune modulatory function of DcR3, we treated feeder-expanded T cells with DcR3, which protected the T cells from activation-induced cell death (Fig. 4G) as well as reduced the potency of their effector functions such as cytotoxic killing (Fig. 4H). Interestingly, while the encoding gene is present in the human genome, there is no orthologue in mice (44). Collectively, these data demonstrate dysregulation of several members of the TNF superfamily in LTβR deficient patients.

Ex vivo co-culture reveals functional B cell activation and differentiation

To further characterize the B-cell compartment in LTβR-deficient patients, we performed B-cell receptor sequencing and analyzed the somatic hypermutation (SHM) rate in the Ig heavy-chain variable regions (IGHV). While the B-cell repertoire diversity was comparable between the patients and HCs, indicating that the process of V-D-J recombination and the ability to generate mature naïve B cells is intact in the patients, we observed a significant reduction in the number of mutations in the IGHV regions and the percentage of individual B-cell clones in which SHM occurred (Fig. 5A, Fig. S14A and Fig. S14B), further corroborating the lack of an efficient GC reaction. Although LTβR signaling is critical for SLO development (17, 27), its direct role in GC formation is less clear (12). Established assays to measure B-cell differentiation utilize cytokine stimulation instead of direct cell-cell interactions. We devised an ex vivo model for studying the ability of different cell types to interact and induce B-cell differentiation to mimic the GC reaction (Fig. S15A-C). We co-cultured different combinations of PBMCs, dendritic cells (DCs) and dermal fibroblasts with or without additional stimulation of a vaccine for measles, mumps and rubella (MMR) in a transwell system. The DCs were differentiated ex vivo from monocytes and then activated with polyinosinic:polycytidylic acid (poly I:C) (Fig. S15B). The dermal fibroblasts were either donor-derived or modified by generating B2M^-/- fibroblasts in order to inhibit HLA-I expression and prevent T-cell mediated apoptosis (Fig. S15C). Following the co-culture, B cells from HCs, despite some level of variability, showed an upregulation of the activation marker CD25 as well as activation-induced cytidine deaminase (AID), the enzyme initiating SHM, (Fig. 5B) (48).

Figure 5. Functional ex vivo B-cell activation in LTβR deficient cells.

(A) The rate of somatic hypermutation (SHM) in the Ig heavy-chain variable regions (IGHV) of the top 200 B-cell clones with productive IGH rearrangements. Each dot represents a unique B-cell clone with the number of mutated nucleotides (nt) plus one on the y axis (logarithmic scale). The percentage of clones with no SHM is shown below each sample in a circular plot. Asterisks represent p-values < 0.001 of Dunnett’s test comparison between each individual patient samples and the aggregate of HC samples. (B) Upregulation of the activation marker CD25 and Activation-induced cytidine deaminase (AID) in CD19⁺ B cells from HCs (n=4) following 7 days of co-culture in various combinations of PBMCs, SCs, DCs and MMR vaccine as indicated in the panel. (C) Differentiation of CD19⁺ B cells from controls (n=4) and patients into CD27⁺ memory or CD38⁺IgD^- GC-like B cells following 12 days of co-culture. Measurements are compared to Day 0. Indicated conditions include the addition of stromal cells (SC), monocyte-derived dendritic cells (DC) or additional treatment with a measles, mumps, and rubella (MMR) vaccine. All conditions were stimulated with B-cell activating factor (BAFF) every other day. Readout was performed using flow cytometry.

Despite the absence of LTβR signaling, patient B cells were able to differentiate into CD27⁺ memory B cells when co-culturing lymphocytes with activated monocyte-derived dendritic cells and stromal cells (Fig. 5C, and Fig. S16). The additional stimulation with the MMR-vaccine induced the differentiation into GC-like (IgD^-CD38⁺) B cells (Fig. 5C, and Fig. S16). Hence, LTβR deficiency does not result in an intrinsic B cell defect.

Taken together, these data suggest that LTβR signaling, although critical for SLO development, becomes redundant for B cell differentiation in a GC-like setting if immune cells and stromal cells are brought into close proximity artificially. Therefore, the observed humoral defects in LTβR-deficient patients may stem from the absence of a conducive environment typically provided by the stromal compartment in SLOs.

Discussion

SLOs play a crucial role in initiating the adaptive immune response and establishing immune memory. Although the formation and organization of SLOs remain active areas of research, LTβR signaling has been identified as a key factor in lymph node development in mice. However, its relevance to humans has been less clear. In this study, we identify LTβR deficiency underlying a previously unknown combined immunodeficiency, characterized by a distinct absence of functional SLOs resulting in hypogammaglobulinemia and low memory B cells. The T-cell compartment alterations are likely contributing to the observed susceptibility to viral infections including b-HPV (21, 49).

While the observed defects in the SLOs are consistent with previously published Ltbr^−/− mouse models (12, 16, 50), other key signs of immune dysregulation observed in mice (18, 42, 51) were not as prominent in LTβR-deficient patients. We speculate that the defect in the SLOs and subsequently aberrated B-cell differentiation and activation, may, paradoxically, protect patients from developing and sustaining autoantibody-producing plasma cells. All three patients have been receiving immunoglobulin substitution, which is a well-established immune replacement therapy for various autoimmune disorders (52), despite its heterogeneity and unclear composition (53). We detected increased levels of the DcR3 protein in serum samples at different time points throughout disease progression in all three patients. DcR3 is a protein which has previously been reported to be anti-inflammatory and anti-apoptotic via blockade of FasL and LIGHT (54), in line with the increased survival and decreased killing activity shown in our study. Intriguingly, DcR3 is not encoded in mice (44, 45). In addition to confirming the immunoregulatory function of DcR3 on human T-lymphocytes, we detected high concentrations of DcR3 in various commercially available IVIG solutions. Therefore, it is possible that the increase of DcR3 in LTβR deficient patients is in part due to the imbalance of several TNF members or the IVIG replacement therapy. Further delineation of a potential role of DcR3 in modulating immune dysregulation will require additional studies in the future.

The TNF superfamily receptors, particularly RANK, CD40, and LTβR, play crucial roles in the formation of the thymic microenvironment and the induction of central tolerance (55). RANK and CD40 signaling pathways are essential for the development of medullary thymic epithelial cells (mTECs) and the expression of the autoimmune regulator (AIRE), a master regulator of ectopic peripheral antigen expression (56). AIRE is vital for negative selection by promoting the expression of peripheral tissue-restricted antigens (TRAs) in mTECs, thereby preventing autoimmunity by ensuring self-tolerance. Mutations in AIRE lead to autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), a monogenic disorder characterized by multiorgan autoimmune destruction with a distinct IFNγ mediated inflammatory signature (57). In contrast, while the LTβR signaling pathway is also necessary for normal thymic architecture and mTEC differentiation, it does not influence AIRE expression or AIRE-dependent TRA expression. Instead, LTβR signaling induces the transcriptional regulator Fezf2, which governs the expression of a distinct set of TRAs (55). In LTβR-deficient mice, Fezf2-dependent TRA expression is reduced in mTECs, highlighting an alternative pathway for TRA expression independent of AIRE, which implicates that LTβR signaling supports the maintenance of thymic stromal cell subsets and chemokine production, playing a broader role in the regulation of the thymic microenvironment. Consistent with this distinct downstream transcriptional impact, the clinical phenotype of LTβR-deficient patients is quite different from that of AIRE deficiency, presenting predominantly as B-cell dysfunction without the IFNγ-mediated chronic inflammation and autoimmunity.

Interestingly, not only Ltbr^−/− mice, but also mice overexpressing the LTβR ligands LT or LIGHT can exhibit signs of autoimmunity (40, 58), likely to be caused by the unregulated formation of tertiary lymphoid organs (TLOs) (59). These are ectopic lymphoid structures that arise temporarily in proximity to sites of chronic inflammation and resemble SLOs in their organization, function and dependence on LTβR signaling (59). Despite their role as gatekeepers and as a favorable prognostic factor in cancer, TLOs are associated with increased tissue damage in autoimmune diseases (59). Thus, LTβR signaling has been a therapeutic target in autoimmune diseases including rheumatoid arthritis or Sjögren’s syndrome, but clinical trials using inhibitors for LTβR (25) or its ligand (26) have been unsuccessful. The results from our ex vivo GC model suggest that once the process of GC formation has been initiated in either SLOs or TLOs, blockade of LTβR may not suffice to stop the ongoing GC reaction and production of autoantibody-secreting plasma cells.

Our study shows that biallelic LOF mutations in LTBR cause an IEI with a predominantly humoral immune deficiency and milder T-cell defects, in contrast to murine models where the role for LTβR signaling in autoimmunity was demonstrated by controlling thymic stroma and intestinal microbiota (60). Notably, and further corroborated by murine data, the non-lymphocyte lineage specific expression of LTβR suggests that allogeneic hematopoietic stem cell transplantation (HSCT) may not provide a curative treatment for human LTβR deficiency (12), whereas anti-infective prophylaxis using IVIG and prophylactic antibiotics is imperative. Our study advances our understanding of LTβR as a critical factor of human immune homeostasis with implications for potential targeted therapies in the context of severe infections and autoimmunity. Whether the observed increase in DcR3 indeed modulates the clinical phenotype warrants further research, and to refine therapeutic strategies, a more profound comprehension of human-specific molecular regulation and tissue-specificity in lymphotoxin signaling is paramount. Another limitation of our study – similar to other studies identifying genetic etiologies of rare diseases - is the relatively low numbers of patients identified with this genetic defect, implying that future studies with additional affected individuals will enable the delineation of the full phenotypic spectrum of disease and potential genotype-phenotype correlations.

Materials and Methods

Study Design

The objective of this study was to investigate the role of LTβR on human immune homeostasis. For this purpose, we performed an array of functional and multiomic experiments on primary material from the patients carrying a germline-encoded homozygous mutation in LTBR, after genetic analysis. Furthermore, we made use of cellular models to investigate the effects of the LTβR mutation on the function of lymphocytes and stromal cells using biochemical and proteomic approaches. This included a self-developed co-culture system of different immune and stromal cells to assess their interaction and capacity to stimulate B-cell differentiation. Control samples were used either from healthy shipment controls or taken from healthy local donors.

Study Oversight

The study was approved by the relevant institutional review boards and performed in accordance with the guidelines of good clinical practice and the current version of the Declaration of Helsinki. Written informed consent was obtained from the patients or the patients’ legal representatives.

Patient And Human Cell Lines

Peripheral blood mononuclear cells (PBMCs) from the patients and healthy controls were isolated by Ficoll gradient. Patient- and healthy-control-derived T cells were expanded by stimulation of PBMCs with irradiated feeder cells, 1μg/mL phytohemagglutinin (PHA, Sigma Aldrich), and 100 IU/mL interleukin-2 (IL-2, Novartis) in RPMI 1640 media containing 5% human serum (IBJB – Inst. Biotechnologies J.BOY, 201021334) and supplemented with 1 mM sodium pyruvate (Thermo Fisher Scientific, 11360039), MEM Non-essential Amino Acid Solution (Sigma Aldrich, M7145), 50 U/mL penicillin, 50 mg/mL streptomycin and 10 mM HEPES. Fibroblasts were isolated from skin biopsies from patients and healthy controls and cultured in DMEM medium containing 10% FCS, 50 U/mL penicillin, 50 mg/mL streptomycin and 10 mM HEPES. All cells were cultured at 37°C in a humidified atmosphere with 5% CO₂.

Whole-Exome Sequencing

Genomic DNA (gDNA) was isolated from peripheral blood samples using commercial extraction kits (P2: DNeasy Blood and Tissue Kit (Qiagen); P3: GenEx Blood (GeneAll)).

Whole-exome sequencing (WES) of P2 involved library preparation and exome enrichment utilizing Nextera Rapid Capture Exome kit (Illumina), followed by 150-bp paired-end sequencing on the Illumina HiSeq3000 system. Sequenced DNA reads were mapped to the human reference genome (GRCh38/hg38 assembly) by means of the Burrows-Wheeler Aligner (BWA) (61). Following variant calling with the Genome Analysis Toolkit (GATK) HaplotypeCaller (62), Variant Effect Predictor (VEP) was used for annotating single-nucleotide variants (SNVs) and small insertions/deletions (63). From the obtained variant calls, non-synonymous (nonsense, missense, small insertions and deletions) as well as splice-region variants (+/-8 bp from the intron/exon boundaries) were then filtered to exclude those with a minor allele frequency > 0,01 in gnomAD v2.1.1 (64).

An in-house database including sequencing data from > 1200 individuals was used to further exclude recurrent variants with an allele frequency > 0,02. The remaining variants were prioritized based on literature research and their Combined Annotation Dependent Depletion (CADD) pathogenicity prediction score (65).

For P3, the WES library was prepared using the Nextera DNA Prep with Enrichment Kit (Illumina) and sequencing was performed with 150-bp paired-end reads on the Illumina NextSeq 550 platform. Data processing, including mapping, variant calling and annotation was conducted with the SEQ Platform v8 (Genomize).

Sanger Sequencing

Isolation and purification of gDNA from the probands and family members of both kindreds was performed from peripheral blood, using the DNeasy Blood and Tissue Kit (Qiagen).

Sanger sequencing was employed for the validation and segregation of the LTBR variants identified through WES in the patients and their family members. Specific primers were designed to amplify the genomic regions encompassing each of the two identified variants. Sanger sequencing primers used for validation and segregation of LTBR variants in the patient and family members are shown in Table S7.

Protein Structure Visualization

The three-dimensional (3D) structural model of wildtype LTβR was obtained from AlphaFold and visualized with PyMOL (Molecular Graphics System, version 2.0 Schrödinger, LLC) (66, 67). MISSENSE3D was used to predict possible effects of p.Arg120Pro on the protein (68, 69).

Crispr/Cas9 Editing Of Human Cells

For the reconstitution experiments, dermal fibroblasts derived from P1 and P3 were edited by CRISPR/Cas9 in order to re-express LTβR. Cells were collected and resuspended in Opti-MEM (Thermo Fisher Scientific, 31985062). For electroporation, 10^6 cells were used per condition. Prior to electroporation, 125 pmol of Cas9 protein (IDT, Alt-R® S.p. Cas9 Nuclease V3) and 150 pmol of single gRNA (sgRNA) were mixed and incubated for 15 min at room temperature. Subsequently, 100 pmol of single-stranded oligodeoxynucleotide (ssODN) template was added to the cells (Table S8). A NEPA21 electroporator (NepaGene) was used for electroporation using the following settings: Poring pulse: V 250, 2.5 ms pulse length, total two pulses with 50 ms interval between the pulses, 10% decay rate with + polarity; Transfer pulse: 20 V, 50 ms pulse length, total 5 pulses with 50 ms interval between the pulses, 40% decay rate with +/- polarity. Next, cells were seeded in DMEM supplemented with 10% FCS. Ten days later, cells were stained with LTβR antibody and sorted using a FACS Aria™ Fusion cell sorter. Positive cells were used for further analysis.

For the co-culture experiments, dermal fibroblasts from a healthy control were edited by CRISPR/Cas9 to knock out B2M to prevent expression of HLA-I. Except for the addition of an ssODN template, the electroporation was performed the same way as described above. Afterwards, cells were seeded in DMEM supplemented with 10% FCS. Ten days later cells were stained with HLA-I antibody and sorted using a FACS Aria™ Fusion cell sorter. Negative cells were used for further analysis. Dermal fibroblasts from P2 were used and processed for TNFR1 knock out the same way.

PBMCs, either fresh or cryopreserved in liquid nitrogen, were used for immunophenotyping. To reduce unspecific antibody binding, the cells were blocked in RPMI containing 10% FCS for at least 1 hour, before surface staining with antibodies (listed in Table S9) for 30 min in the dark at 4°C. Stained cells were acquired with an LSRFortessa™ (BD Biosciences) or FACSymphony™ (BD Biosciences). FlowJo™ v.10 was used to analyze the data and Prism v.8 (GraphPad) was used to produce graphs. For intracellular staining, the cells were permeabilized after surface staining with BD 1x Perm/Wash buffer before additional staining with intracellular antibodies for 30 min in the dark at 4°C. The complete list of antibodies used in this study is shown in Table S9.

T-Cell Activation And Proliferation Assay

Feeder-expanded T cells from patients and healthy controls were stained with violet proliferation dye (VPD450, BD Biosciences) and seeded on 96-U-shaped plates at 400,000 cells/ well in 100 μl RPMI 1640 supplemented with 10% FCS. Cells were stimulated with either 1 μg/mL soluble anti-CD3 (sCD3, clone: OKT3), a combination of 1 μg/mL sCD3 with 1 μg/mL soluble anti CD28 (sCD3/CD28, eBioscience), 5 μg/mL phytohemagglutinin (PHA, Peprotech), CD3/CD28 DynaBeads (Beads, Invitrogen) or left untreated. Activation was assessed after 24 hours using flow cytometry. Proliferation was measured after 4 days of stimulation by dilution of the VPD450.

T-Regulatory Differentiation Assay

The T-regulatory differentiation assay was used according to the to the manufacturer’s protocol (CellXVivo Human Treg Cell Differentiation Kit #CDK006) with small adaptations. In brief, cryopreserved PBMCs from healthy controls and patients were thawed, and naïve T-cells were isolated by magnetic bead separation (Miltenyi 130-097-095). From each donor, 50,000 cells were either resuspended with RPMI 1640 complete supplemented with 10% FCS or treated using Human Treg Differentiation Media as prepared by the manufacturer’s instructions. Treated cells were then seeded into a 96-well ELISA plate that had been coated with anti-CD3 (clone: OKT3) for 24 hours, while the untreated cells were added to uncoated wells. Differentiation of T-regulatory cells was measured following 5 days of cultivation using flow cytometry.

T-Helper Cytokine Production Assay

Cryopreserved PBMCs from patients and healthy controls were thawed and seeded on 96-U-shaped plates at 300,000-500,000 cells/well in RPMI 1640 supplemented with 10% FCS. Cells were stimulated with 200nM Phorbol-12-myristat-13-acetat (PMA, Sigma Aldrich) and 1μg/mL ionomycin (Sigma Aldrich) or left untreated for a total of 6 hours. After 1 hour of stimulation Brefeldin-A (BioLegend) was added to the cells. Cells were stained with surface antibodies before fixation and stained with intracellular cytokine antibodies. Cells were analyzed using flow cytometry.

B-Cell Activation, Proliferation And Class Switch Assay

Cryopreserved PBMCs from P1 and P2 and healthy controls were thawed and stained with violet proliferation dye (VPD450, BD Biosciences) before being seeded on 96-U-shaped plates at 400,000 cells/well in 100 μl RPMI 1640 supplemented with 10% FCS. Cells were stimulated with either 100 ng/mL IL-4 (Peprotech) + 200 ng/mL CD40L (Peprotech), 20 ng/mL IL-21 (Peprotech) + 200 ng/mL CD40L, 50 nM CpG (InvivoGen) or left untreated. Activation was assessed after 24 hours using flow cytometry. Proliferation was measured after 6 days of stimulation by dilution of the VPD450. Class Switch was measured after 5 days of stimulation by CD19⁺ B cells expressing IgA or IgG.

B-Cell Differentiation Assay

Cryopreserved PBMCs from patients and healthy controls were thawed and B cells were isolated by magnetic bead separation (Miltenyi 130-101-638). Cells were seeded in 96-U-shaped plates at 50,000 cells/well in 100 μl RPMI 1640 supplemented with 10% FCS. Cells were stimulated with a combination of 5 μg unconjugated goat anti-human F(ab)2 fragments (Jacson Labs), 50 ng/mL IL-21 (Peprotech), 1 μg/mL CD40L (Peprotech) and 2.5 μg/mL CpG (Peprotech) or left untreated. Differentiation was measured following 6 days of stimulation using flow cytometry.

Stimulation Of Dermal Fibroblasts For Dcr3 Production

Dermal Fibroblasts from healthy controls, patients and were from P2 where TNFR1 was knocked out using the CIRSPR/Cas9 system, were seeded at 15,000 cells/well in 200μl in 48-well plates. After 24h hours cells were stimulated with 100ng/mL human TNFα (Miltenyi), 200ng/mL human TNF-β/LTα3 (Peprotech), 200ng/mL human lymphotoxin alpha1/beta2 (LTα1β2, R&D Research) or 200ng/mL human lymphotoxin alpha2/beta1 (LTα2β, R&D Research) or left untreated. After 24 hours supernatant was collected and used for DcR3 analysis using ELISA (R&D Systems). Results from two independent experiments were averaged, before each sample was normalized to its untreated result, followed by pooling of the data of healthy controls and patients.

Effect Of Dcr3 On Activation Induced Cell Death (Aicd) In T Cells

96-U-shaped plates were coated with 2 μg/mL CD3 (clone OKT3) overnight at 4°C, before being washed with PBS. Following 10 days of feeder stimulation with IL-2 (Novartis) and PHA, expanded T cells from healthy controls and from P2 were seeded at 100,000 cells/well and treated with 3.75 μg/mL DcR3 (MedChemExpress) or left untreated. Following 24 hours of stimulation, cells were harvested, and apoptosis was assessed via staining with Annexin V-APC (BD Biosciences) and propidium iodide (PI, BioLegend) and readout was performed using flow cytometry.

Effect Of Dcr3 On T-Cell Killing Function

GFP+ P815 target cells were treated with 1 μg/mL of anti-CD3 (clone: OKT3) antibody for 1 hour or left untreated. Expanded T cells from two healthy controls previously stimulated for 10 days with feeder cells, IL-2 (Novartis) and PHA were assessed for CD4⁺ and CD8⁺ ratio before treatment for 1 hour with 6.25 μg/mL DcR3 (MedChemExpress), 37.5 μg/mL DcR3 or left untreated. P815 cells were washed before being co-cultured with the expanded T cells at a 1:1 ratio in 96-U-shaped plates. 0.2 μg/mL Aphidicolin (Sigma Aldrich) was added to inhibit proliferation of the cells. Following 6 hours of co-culture cells were harvested and the killing function of T cells was estimated by measuring the amount of 7-AAD negative (BD Biosciences) P815 cells by flow cytometry.

Co-Culture For B-Cell Differentiation

For the co-culture, different combinations of PBMCs, dendritic cells (DCs) and fibroblasts were combined with or without additional stimulation of a vaccine for measles, mumps and rubella (MMR; Merck Sharp und Dohme). PBMCs from patients and healthy controls were thawed and CD14⁺ monocytes isolated using CD14 MicroBeads (Miltenyi). Monocytes were then differentiated to DCs over 6 days by adding 1000 IU/mL IL-4 (Peprotech) and 1000 IU/mL GM-CSF (Peprotech). DCs were stimulated by adding 10 μg/mL polyinosinic-polycytidylic acid (poly I:C, InvivoGen) for 24 hours. CRISPR-Cas9-edited B2M knockout fibroblasts or patient derived fibroblasts were seeded at 5,000 cells/well into a 24-well transwell. The next day, the donor-specific PBMCs were added at 300,000 cells/well together with donor-specific DCs at 30,000 cells/well with or without addition of 5 μl of MMR per well. 1 ml of RPMI 1640 supplemented with 10% FCS was added in the transwell below. Cells were stimulated with 50 ng/mL B-cell activating factor (BAFF, Peprotech) every other day. Cells were harvested on day 7 for activation readout or on day 12 for differentiation readout using flow cytometry.

NF-κB Stimulation In Fibroblasts

For canonical NF-κB activation, dermal fibroblasts were stimulated with 20 ng/mL human TNFα (Miltenyi) for 10-60 minutes followed by immunoblot analysis. For non-canonical NF-κB activation, dermal fibroblasts were stimulated with 200 ng/mL human lymphotoxin alpha1/beta2 (LTα1β2, R&D Research) for 6 hours followed by immunoblot analysis.

Immunoblot

Cell lysates of healthy control and patient fibroblasts were prepared in RIPA buffer containing protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific). Protein concentrations were determined using the DC Protein Assay Kit II (Bio-Rad), and 20 μg of protein were used and resolved by reducing SDS-polyacrylamide gel electrophoresis with 4-15% mini-protean TGX precast protein gels, followed by transfer to PVDF membranes using the Trans-Blot Turbo RTA Mini 0.45 μm LF PVDF Transfer Kit and the Trans-Blot Turbo transfer system (Bio-Rad). Then, membranes were blocked in 5% BSA solution for 1 hour at room temperature and subsequent incubation with primary antibodies NF-κB2 p100/p52 (Cell Signaling) and HSP90α/ß (Santa Cruz Biotechnology) or with Phospho-NF-κB p65 (Cell Signaling), IκBα (Cell Signaling) and HSP90α/ß (Santa Cruz Biotechnology) overnight at 4°C. Subsequently, membranes were washed three times in TBST and incubated with peroxidase-conjugated secondary antibodies for 1 hour at room temperature followed by three times washing in TBST and visualized using chemiluminescence with the ECL Prime Western Blot Detection Reagent (Cytiva) and the ChemiDoc MP Imaging System (Bio-Rad).

Enzyme-Linked Immunosorbent Assays (Elisa)

LTB, TNF-α, CXCL13/BLC, DcR3/TNFRSF6B, Fas Ligand/TNFSF6 and LIGHT/TNFSF14 levels were detected by enzyme-linked immunosorbent assay (ELISA; LTB: Cat. A312081, Antibodies.com; TNF-α: Cat. Nr. 10737663, Thermo Fisher Scientific; CXCL13/BLC: Cat. Nr. 15444963, Thermo Fisher Scientific; DcR3: Cat. nr. 15405163, Thermo Fisher Scientific; Fas Ligand: Cat. nr. DY126, DuoSet kit R&D Systems; LIGHT: Cat. nr. DY664, DuoSet kit, R&D Systems) in healthy control and patient serum samples according to the manufacturer´s instructions. 50 μl of each serum sample in duplicates was used in all assays.

Single-Cell Rna Sequencing

Samples from P1 and P2 were analyzed along with four healthy controls as previously described (70). In brief, cryopreserved PBMCs from the two patients and four healthy controls were thawed, washed in RPMI media and resuspended in sterile PBS with 0.04% bovine serum albumin (BSA). A single-cell suspension was obtained by passing one million cells into a 5 mL FACS tube through a cell strainer and sorting for the live lymphocytes and monocytes based on the forward and side scatter using the FACSAria™ Fusion (BD). Single-cell RNA-seq was then performed on the live samples using the Chromium Single Cell Controller and Chromium Next GEM Single Cell 5’ Kit v2 (10X Genomics, Pleasanton, CA), according to the manufacturer’s protocol. T-cell receptor (TCR) sequences were enriched at the cDNA stage using the respective reagents, in accordance with the instructions of the VDJ Kit workflow by 10X Genomics. Sequencing was performed using the Illumina NovaSeq platform in the 75 bp paired-end configuration.

CellRanger v5.0.1 software (10x Genomics) was used for demultiplexing and alignment to the GRCh38-2020-A human reference transcriptome. The R statistics software was used to analyze the processed data. Briefly, CellRanger outputs (filtered count matrices) were further filtered to exclude cells with more than 15% mitochondrial counts or with numbers of detected genes being either less than 300 or unusually high (per-sample z-score > 2.5). Cell types were annotated using Seurat (v 4.1.0) with sctransform (71) normalization (v.0.3.3) and Azimuth (72) (v 0.4.3). The reference used was Human PBMC annotation level 2. Cells with annotation score or mapping score less than 0.5 were excluded from further analysis. For visualization, the Azimuth reference space (UMAP) was used and cell types with less than 20 cells annotated (cDC2, pDC, CD16 Mono, Platelet, HSPC, ILC, dnT) were excluded. Clonality analysis of the TCR repertoires was based on cells that were assigned exactly one alpha chain CDR3 and one beta chain CDR3. A clonotype was defined as a unique combination of these CDR3 motifs and clonality for patients and cell types was calculated as 1 minus normalized entropy as done in other studies (73). Single-cell RNA-seq differential expression (DE) testing of patients vs healthy controls was done using edgeR (74) (v 3.36, default parameters, exact test) and pseudobulk profiles (aggregated counts per patient) to minimize false discoveries (75). Each test was filtered to exclude genes with FDR-adjusted P-value > 0.05; the remaining genes were ordered by p-value. Several tests were performed focusing on different cell populations each. Tests were performed within CD8⁺ T cells (aggregating CD8 Naïve, CD8 TCM, CD8 TEM) and within the individual CD8⁺ cell types. Expression heatmaps show normalized expression (DESeq2 vst function after aggregating counts of all cells per group) z-scored per gene and cropped to the range -2 to 2. Genes selected for display in the heatmaps had to have p-value rank <= 45 and logFC rank <= 45 (grouped by fold-change direction) in one of the CD8 cell type differential expression tests. Following differential gene expression analysis, we performed gene set enrichment analysis using hypergeometric tests. For the tests, we considered all the genes retained by edgeR as background, while the DE genes (FDR <= 0.05) were considered the genes of interest. The following genes defined the “exhaustion” gene set: PDCD1, CTLA4, NFATC1, SPRY2, BATF, VHL, FOXO1, FOXP1, LAG3, CD244, CD160, HAVCR2, TRAF1, TNFRSF9, IL10RA, IL10, PRDM1, STAT3, IFNA1, IFNB1, IL21, CXCR5, SOCS3, GATA3, IKZF2, BCL6, BCL2, TBX21, EOMES. A significant enrichment for exhaustion was not observed in any of the differential expression test results (p-values > 0.1).

B- And T-Cell Receptor Sequencing

For B-cell receptor (BCR) and T-cell receptor (TCR) sequencing, DNA was isolated from PBMCs or whole blood of patients and healthy controls. DNA used for BCR sequencing of P1 and P2 was isolated from B-cell enriched samples using magnetic MACS separation as described previously in the B-cell differentiation protocol. Sequencing was performed using protocols and primers standardized by the EuroClonality-NGS Working Group and sequenced with Illumina MiSeq v3 600-sequencing kit (MS-102-3003) with 20% PhiX v3 Control library (FC-110-3001; both Illumina, San Diego, CA, USA) following the manufacturer’s instructions (76, 77). Demultiplexed sequencing data were analyzed using the ARResT/Interrogate platform to annotate individual clones for further processing with R (78, 79). For the analysis of clonality, only productively rearranged IGH and TCRB clones were included in the analysis, and only the top 1000 clones by frequency were included for the figure plots and diversity calculations. For the analysis of the somatic hypermutation (SHM) rate, all productively rearranged IGH clones with unique amino acid sequence in the complementarity determining region 3 (CDR3) and with a coverage of at least 5 reads in the sample were used, and its representative sequence with the highest number of reads was used for alignment. The alignment to the IGHV gene reference sequences was performed using the IMGT/HighV-QUEST platform and resulting files were processed by custom R code to generate statistics and plots (80, 81). The number of nucleotide mutations was calculated for each clone with a productive IGH rearrangement in the CDR1, CDR2, FR2 and FR3 regions of the rearrangement. For the figure plots and underlying statistics, the top 200 clones by their frequency were analyzed for each sample. A detailed analysis of the IGHV3-23 gene was performed for all clones with a productive rearrangement of this IGHV gene and a mutational rate was calculated for each amino acid position based on the alignment to the IMGT reference. A mutational rate is shown as a median per position for all four healthy controls with an error bar corresponding to the 75^th percentile and individual values per position for all patients.

Statistical analysis

For individual comparisons of independent groups, the Student t-test was performed. Welch’s correction was used if the two groups had unequal variances or sample sizes. For multiple comparisons, a one or 2-way analysis of variance (ANOVA) was applied, followed by Bonferroni post hoc test to correct for multiple comparison. In case of multiple comparisons towards one dataset (e.g. untreated) Dunnet’s post-hoc test was used instead. Data graphs and analyses were made using PRISM software (GraphPad Software Inc) and error bars display the standard deviations.

One Sentence Summary.

Mutations in LTBR resulting in a stromal defect leading to dysfunctional secondary lymphoid organs and combined immunodeficiency.

Data and materials availability

Raw sequencing reads are deposited in the European Genome-Phenome Archive (accession numbers EGAS00001007271 [BCR/TCRseq], EGAS00001007271 [scRNAseq]. These data are available via controlled access to safeguard patient privacy. R code used for the analysis of single-cell RNA-sequencing data is available on GitHub at https://github.com/cancerbits/ransmayr2024_ltbr and for the analysis of the BCR and TCR sequencing results at https://github.com/msvtnCCRI/ransmayr2024_ltbr_igtr. All other data needed to support the conclusions of the paper are present in the paper or the Supplementary Materials. Tabulated underlying data for all figures can be found in supplementary data file S1.