Immunodeficiency in SIOD Patients is Hallmarked by Profound T cell Lymphopenia

In this study, we explored the immune characteristics of 4 patients with genetically verified SIOD (3 male, 1 female, age at sampling 9.7 ± 5.65years) followed at the Department of Pediatrics and Department of Internal Medicine, Motol University Hospital in Prague, Czech Republic. Patient characteristics are shown in Table 1 and were discussed in detail in previous publications [3]. Patients presented at early childhood (2.3 ± 1.01years) with typical clinical features - disproportionate short stature, facial dysmorphism and steroid resistant proteinuric nephropathy. Average time to end-stage kidney disease (ESKD) was 1.5 ± 0.44years. Peritoneal dialysis was started in all children. Patient 1 died at the age of 20 years of cardiorespiratory failure as a result of progressive bacterial peritonitis and patient 2 died at the age of 4 years and 8 months of cardiorespiratory failure of uncertain etiology.

Table 1 – Clinical and routine laboratory data of the individual patients, normal ranges for laboratory parameters in round brackets

All patients had marked CD4 and CD8 T cell lymphopenia, which necessitated antibiotic prophylaxis with trimethoprim/sulfamethoxazole in 3 out of four patients. B cells were less affected, with a reduction of total B cell counts in only 1 out of 4 patients. However, immunoglobulin replacement was necessary in 2 patients due to stark reduction of serum IgG levels - notably, the eldest patient (P1) with only a discrete reduction of IgG bears a missense rather than a termination or compound missense/termination mutation. Parents of P2, whose IgG levels were severely decreased, did not approve SCIG therapy after adverse effects during the first dose. The development of B cells was affected in 2 out of 3 tested patients leading to a reduction in mature class-switched memory cell stages.

SIOD T Cells are Skewed Towards a Memory Phenotype With a Strong Activation And Th1 Bias

To further investigate the T-cell phenotype beyond the low overall counts, we performed developmental T cell phenotyping in all patients using in vitro cell culture and spectral cytometry with functional readouts.

Therein, we observed significant differences characterized chiefly by a significantly decreased population of recent thymic emigrants (RTEs, CD4 + 31 + 45RA + 62 L+, Welch’s t-test p = 0.01) and mature naïve (CD45RA + 62 L+, 2-way ANOVA p = 0.004) T cell counts, with a relative increase of effector memory cells (CD45RA-62 L-, p < 0.0001) (Fig. 1A). Patient T cells displayed a significant loss of naïve and stemness markers CD27, CD28 and the transcription factor TCF1 (p = 0.0003, 0.0004 and < 0.0001, respectively) (Fig. 1B, Supplementary Fig. 2), and a concurrent increase of exhaustion-associated inhibitory receptors PD-1 (p = 0.0002) and Tim3 (p = 0.0005), the marker of cellular replicative senescence CD57 (p = 0.0004) and the pro-apoptotic receptor CD95/Fas (p > 0.0001) (Fig. 1C, Supplementary Fig. 2). The previously described decreased expression of CD127 [9], the surface receptor for IL-7 expressed mainly in naïve T cells, was also present in our data (p = 0.0119), but was mild and did not reach statistical significance in any individual T cell subset.

Fig. 1

Memory populations and phenotype of SIOD patient T cells. RTE, naive, effector memory and Th1 T cells ex vivo (A). Expression of maturation (B), exhaustion/senescence (C), activation/proliferation (D) and functional readout markers (E) after 72 h culture. Y axis shows percentage of cells positive for a marker from population specified on X axis. Statistical significance was determined using two way ANOVA (significance for the “cohort“ variable indicated by asterisk in the plot title *p < 0.05, **p < 0.01 ***p < 0.001, ns = not significant) with multiple comparisons test corrected by Two-stage Benjamini, Krieger and Yekutieli method. FDR adjusted p-value of 0.05 was considered significant: *q < 0.05, **q < 0.01 ***q < 0.001

This highly differentiated phenotype was driven through increased proliferation measured by the expression of nuclear protein Ki-67 (p < 0.0001) (Fig. 1D, Supplementary Fig. 2). The persistent stimulation towards proliferation was reflected in T cell activation, with increased HLA-DR (p < 0.0001) and CD38 expression (p = 0.02), whereas the early activation marker CD69 was decreased (p < 0.0001).

Interestingly, within memory CD4 T cells we also observed a significant increase of the pro-inflammatory CXCR3 + CCR6- T helper 1 fraction (1-way ANOVA p < 0.0001) (Fig. 1E), and increased IFN-γ and IL-2 production by patient T cells (2-way ANOVA p = 0.0326, 0.0453, respectively). Similarly, cytotoxic CD8 T cells produced highly increased levels of granzyme B (p < 0.0001).

To exclude the effect of peritoneal dialysis on T cell maturation and activation we also analyzed 3 patients on peritoneal dialysis necessitated by other causes not associated with immunodeficiency and saw no difference in their T cell phenotype compared to healthy donors (Supplementary Fig. 1), suggesting that the reduction of naïve stages, upregulation of PD-1, activation and skew towards Th1 is an innate characteristic of SIOD.

Thus, we show that T cells in SIOD are fewer, skewed away from naïve and towards memory stages, with activated and exhausted phenotype and amplified proinflammatory features, especially with augmented production of IFN-γ and cytotoxicity.

Interleukin-7 does not Cause Significant Difference in the T cell pool despite Lower CD127 Expression

To assess whether the T cells in SIOD respond differently to IL-7, we incubated HD and SIOD PBMCs for 72 h with/without 10ng/ml IL-7. In HD-derived T cells, IL-7 culture caused a slight but insignificant downregulation of CD127 (Fig. 2B) and a concomitant decrease of naïve T cells, with corresponding increase in the terminally differentiated TEMRA subpopulation (Fig. 2A). SIOD-derived T cells, on the other hand, failed to respond in similar fashion. No significant changes were observed in other phenotypic and functional features in either cohort (ratio paired t-tests for each marker native/IL-7 treated not significant) (Fig. 2B).

Fig. 2

Effect of IL-7 on T cell phenotype. Proportion of naive, effector/central memory and TEMRA subpopulations from the CD4 and CD8 T cells respectively, in healthy controls and SIOD patients, for both IL-7 enriched and non-enriched culture media (A). Comparison of the CD4 and CD8 T cell phenotype in cells treated by IL-7, shown as log2 fold change of the percentage of cells positive for markers specified on the X axis compared to IL-7 non-treated cells (B)

SIOD T Cells are more Prone to Apoptosis and fail to Repair DNA Damage

The changes to T cell phenotype in SIOD patients suggest a significant impairment of physiological T cell development. As the SMARCAL1 protein seems to be associated with DNA damage repair [13, 14], we hypothesize that SIOD patient T cells may be susceptible to DNA damage. This, in turn, would result in T cell apoptosis during thymopoiesis, in particular in the VDJ recombination stages, ultimately leading to low thymic output and hyperactive homeostatic proliferation leading to premature T cell aging.

To test this theory, we irradiated fresh blood by UV light and assessed the amount of double-strand breaks (DSB) compared to HDs (26,58 ± 10,68 years) over time by measuring the presence of γH2X, a phosphorylated histone complex component which signals the presence of DBS to DNA repair proteins [15].

We found a significant increase in spontaneous DSB in SIOD patients (p = 0.0216) and a failure to repair the DSBs induced by UV irradiation (Fig. 3A). While both patients and controls had similar amounts of DSBs 1 h after irradiation, the HDs repaired the damage within 24 h while the SIOD patients failed to do so (p = 0.0063). This increased persistence of DNA damage resulted in increased apoptosis, where after 24 h an average of 15.3% of SIOD T cells were apoptotic, in contrast to 5.9% of cells in the HD group (p-value 0.0167) (Fig. 3B). The difference was exacerbated by UV irradiation with significant difference of apoptotic cells after 1 (p = 0.029) and 6 (p = 0.045) hours from UV irradiation.

Fig. 3

Response to DNA damage. Comparison of T cells positive for γH2X, indicator of double strand breaks, emerging spontaneously (left) or caused by UV induced DNA damage (A), after 1 and 24 h since intervention. Comparison of Fluorochrome-Labeled Inhibitors of Caspases (FLICA) positive T cells, measuring spontaneous or UV induced apoptosis (B), after 1, 6 and 24 h since intervention. Statistical significance was determined using unpaired t-test with multiple comparison corrected by Holm-Šídák method. *q < 0.05, **q < 0.01 ***q < 0.001

RNA Sequencing Reveals Preserved but Unique Transcriptomic Response to DNA Damage

The changes in T cell differentiation are rooted on transcriptional level. Assessing RNA expression in PBMCs using the 581 gene-wide nanoString Human Immunology kit, we noted 65 differentially expressed genes, including decreased expression of IL7R, TCF7, CD27 and CD5 (Fig. 4A), alongside increased expression of the anti-apoptotic LGALS3, and the costimulatory CD70 and CD8 T cell memory-formation promoting BATF3 [16]. Increased TNF mRNA expression in PBMCs yet unincreased TNF-α production by T cells in cell culture suggests that other cells, such as monocytes, might be another potent source of TNF-α in SIOD patients. With the exception of TNF, the transcriptomic analysis and flow cytometry produced equal results for the markers measured by both methods (Supplementary Fig. 3).

Fig. 4

Transcriptomic features of SIOD PBMCs and their response to DNA damage. Volcano plots showing differential expression of genes in SIOD × HD PBMCs ex vivo (A) and after UV irradiation (B). Log2FC of genes is shown on X axis, positive values correspond to upregulation in SIOD. Y axis represents Benjamini-Hochberg FDR q-value as negative logarithm with significantly differentially expressed genes positioned above the horizontal line, capped at 10− 5 (for 4 A) or 10− 8 (for 4B). Principal Component Analysis of differentially expressed genes with 30 top contributing genes shown (C). Enrichment of gene sets containing genes up- or downregulated in HD after UV light irradiation, as expressed by irradiated SIOD PBMCs (D). Bubble plot of genes with opposite direction of differential expression after UV irradiation in HD and SIOD – upregulated in HD but downregulated in SIOD (E) or downregulated in HD but upregulated in SIOD (F)

Principal component analysis (Fig. 4C) of all assessed genes shows a clear separation between the cohorts both in native state and after UV irradiation. As expected, genes with the strongest contribution to native HD cell phenotype include IL7R and CD28, hallmarks of T cell stemness and naivety. On the other hand, native SIOD cells are characterized by the expression of leukocyte immunoglobulin-like receptor subfamily genes, including LILRA5, which contributes to stimulation of TNF-α and other proinflammatory molecules production [17], LILRB1 and LILRB2, which bind MHC-I and provide negative signals to T cells, thus inhibiting CD8 T cells [18], and PECAM1 coding the adhesion molecule PECAM-1/CD31, another potent stimulator of TNF-α production in monocytes [19]. TNF itself was also strongly contributing to the native SIOD gene expression profile.

After UV irradiation, healthy cells upregulated 142 and downregulated 114 genes (Supplementary Figs. 4, 5 and 6). We used these genes to construct a “DNA-repair signature” and analyzed whether SIOD cells utilize similar transcriptomic changes to resolve UV-induced DNA damage. Indeed, SIOD cells strongly enriched the same signature after UV irradiation, suggesting that while some intrinsic differences between the groups persist, the response to DNA damage is largely similar (Fig. 4D). Despite the similar overall pattern (see also vectors in Fig. 4C), the number of genes expressed differentially between HD and SIOD increased from 65 to 154 after UV irradiation, including the novel upregulation of proinflammatory genes such as IL6, S100A8, S100A9, which were not differentially expressed in native cells (Fig. 4B).

We also identified 15 genes with opposite directions of response to UV irradiation between the two groups, i.e. genes upregulated by irradiated healthy cells yet downregulated by irradiated SIOD cells (Fig. 4E), or vice versa (Fig. 4F). Notably, SIOD PBMCs further upregulated (whereas HD PBMCs downregulated) a host of anti-apoptotic genes such as LGALS3 (galectin 3) [20], KIT, BCL6 or ARG2, possibly in an attempt to abrogate the DNA damage. Additionally, irradiated SIOD cells upregulated the expression of IL1RAP, which increases the (similarly increased) IL-6 production [21] and amplifies response to the (also increased) KIT oncogene [22].

Overall, transcriptomic analysis showed downregulation of genes related to T cell naivety and upregulation of genes associated with proinflammatory response, especially of innate immune cells. Patient cells responded in an orchestrated fashion to DNA damage, but interestingly, unlike cells derived from healthy donors, they responded with further augmentation of proinflammatory and anti-apoptotic genes.