We first confirmed previous findings showing the strong association of splenomegaly with granulomatous disease and autoimmune disorders [6, 10, 14]. We also found a significant link with lymphoid hyperplasia and portal hypertension due to hepatic nodular hyperplasia observed in 9 patients with splenomegaly. Patients with splenomegaly tended to have lower percentages of total B cells but the distribution of B-cell subsets was significantly perturbed with fewer smBs and marginal zone B cells, and higher percentages of naïve B cells, as observed in the EUROclass trial [7]. They also suffered from more profound hypogammaglobulinemia for all Ig classes, which correlated with their rarified smBs. Splenomegaly was associated with significantly fewer CD4+ T lymphocytes and Tregs, but their CD4+/CD8+ ratios did not differ from those of patients without splenomegaly. CD4+ and CD8+ T-cell subsets were significantly more activated when splenomegaly was present, while peripheral NK and DC counts were significantly lower.
We then analyzed the impact of the splenectomy on the distributions of circulating-lymphocyte subsets hypothesizing that splenectomy might explain some of their phenotypic characteristics in CVID patients. Total blood-lymphocyte counts increased dramatically post-splenectomy, predominantly attributable to expansion of B cells and CD8+ T cells.
Our main finding is the significant lymphopenia in patients with splenomegaly that normalized post-splenectomy. That observation is fundamental because lymphopenia is an exclusion criterion for CVID diagnosis, leading patients to be diagnosed with “Late onset combined immunodeficiency” (LOCID) [15]. Among our 31 patients with splenomegaly, 18 had CD4+ T-cell counts < 200/mm3, thereby satisfying the LOCID definition of a combined deficiency. However, post-splenectomy, the 4 with CD4+ T lymphopenia below that threshold all saw their lymphopenia corrected, thereby excluding the diagnosis of combined deficiency and returning to the definition of CVID. Patients thought to have a combined deficiency because of lymphopenia might not have true lymphopenia because of splenomegaly. With this knowledge, we believe that CVID-diagnosis should still be attributed for patients with splenomegaly despite lymphopenia.
However, the B-cell distribution remained unchanged by splenectomy, with persistence of fewer smBs, suggesting that this deficiency is intrinsic to CVID and not related to smB-trapping in the spleen. In contrast, splenectomy led to notable changes of the circulating T-cell composition, with inversion of the CD4+/CD8+ ratio that usually declines post-splenectomy in normal individuals [16], suggesting that CD8+ T-cell–compartment amplification is a characteristic specific to CVID patients with splenomegaly. That notion was confirmed by the persistent lymphocyte expansion over time (Fig. 3) and is therefore not influenced by intercurrent factors such as infections. Our results highlight the importance of this CD8+ amplification in such patients, which could be explained as a homing effect to the spleen and/or possible chronic viral replication, which in turn could induce T-cell expansions.
We previously showed that CVID patients, notably those with complications, had oligoclonal expansions of their CD8+ T-cells—not resulting from acute clinical infectious events—reflecting a stable chronic phenotypic T-cell pattern, possibly in relationship with chronic viral replication, facilitated by the immunodepression [11]. However, this oligoclonal expansion was mainly observed in CVID patients with splenomegaly, a finding that suggests a trapping process in the spleen when it is enlarged. Pertinently, that the percentages of the different CD8+ T-cell subsets were unaffected by splenectomy compared to pre-splenectomy values is a strong argument for a homing phenomenon.
To explore that hypothesis, we examined the cellular compositions of spleens excised from 8 CVID patients and compared them to those of 23 controls. The distributions of lymphocyte subsets in normal spleens differed markedly from those in peripheral blood. Langeveld et al. analyzed the splenic lymphocyte compositions of 16 organ-transplant donors [17]. Compared to their peripheral blood, their spleens contained significantly more B cells than T cells, more CD8+ T cells and less CD4+ T cells with an inverse CD4+/CD8+ ratio, a higher percentage of activated T cells and more CD8+ cytotoxic T cells. The same cellular distribution pattern was observed in CVID patients’ spleens but highly amplified; compared to normal spleens, although the CVID spleens contained few B cells, CD8+ activated-T cells were expanded, suggesting trapping of these cells in these patients’ spleens.
Further studies are needed to elucidate why CD8+ T cells are retained in the spleen for longer periods in CVID patients. Their egress from the spleen into blood requires a sphingosine-1-phosphate (S1P) concentration differential at the site of transmigration [18], i.e., high in blood and lymph compared to the interstitial fluid of lymphoid organs. Moreover, S1P receptor-1 (S1PR1) expression on lymphocytes was found to be critical for lymphocyte egress from the thymus and secondary lymphoid organs. Egress modulation is regulated according to the S1P-concentration–S1PR1-expression balance. It is under the influence of several markers, e.g., the activation marker CD69, which limits exit during an immune response by inhibiting the S1PR1 function [19], or factors controlling S1PR1 transcriptional regulation. In this context, CD69 and S1PR1 expressions on T lymphocytes of CVID patients with splenomegaly warrant further investigation. Similarly, another hypothesis to examine would be the influence of chemokines on T-cell–migratory activity in the spleen, particularly the stromal cell-derived factor-1 (SDF-1)/CXCR4 (the α-chemokine receptor specific to SDF-1) axis, whose dysregulations have been implicated in idiopathic CD4+ T-cell lymphocytopenia [20]. An additional avenue of exploration would be potential T-lymphocyte recruitment via aberrant SDF-1 expression by the spleens of CVID patients that would act as a transient reservoir.
Our study obviously has some limitations. First, our sample size is small. However, recourse to splenectomy is not common for CVID patients because of its associated increased infectious risk; thus, we have a meaningful number of splenectomized patients whose phenotypes could be compared before vs. several times post-splenectomy, which has never been reported previously. Second, the numbers of phenotypes obtained for each patient and their frequencies are quite heterogeneous. Nonetheless, phenotyping was repeated at least 3 times over the 3 years after splenectomy for the 8 splenectomized patients followed post-surgery, supporting the good reliability of those results. Finally, the lack of pre-splenectomy phenotyping for 10 of the 21 splenectomized patients represents a data-analysis bias.
In conclusion, splenectomizing CVID patients with splenomegaly restores the absolute circulating lymphocyte count, suggesting that these patients do not have a true combined deficiency and meet CVID criteria.