Patients undergoing HSCT often experience late recovery of B cell numbers leading to a defect of B cell mediated immunity. Generally, B cell numbers recover to normal counts within 12 months after HSCT (25), although complete recovery may take up to 2 years. In the first few months, very few circulating B cells have been observed (25, 26) and within 1–2 years following HSCT, B cell numbers reach levels exceeding normal adult individual ones followed by gradual decline, similarly to the normal ontogeny in young children (26). First B cells emerging into the periphery are CD19⁺CD21^lowCD38^high transitional B cells, which subsequently decrease in percentages while mature CD19⁺CD21^highCD27⁻ naive B cells are being replenished (1, 23). Transitional B cells were first described as CD24^highCD38^high (23). Later on, another marker of transitional B cells was identified, distinguishing between T1 and T2 transitional cells. T1 cells were reported as CD21^low and described as the first B cell population emigrating from the BM, which subsequently differentiate toward CD21⁺, T2 phenotype and serve as precursors of the CD19⁺CD21^highCD27⁻ naive B cell pool in PB and tissues (27). Complete reconstitution of the B cell compartment includes the recovery of both CD19⁺CD21^highCD27⁻ naive and CD19⁺CD27⁺ memory B cells. Reconstitution of memory B cells occurs upon environmental or vaccine-based antigen exposure and requires CD4⁺ T cell help (28). Complete CD19⁺CD27⁺ memory B cell development may take up to 5 years after HSCT (26). In the study by Corre and colleagues, numbers of CD19⁺CD21^highCD27⁻ naive B cells normalized by 6 months and reached above normal values around 24 months after myeloablative conditioning for allogeneic HSCT (29). CD19⁺CD27⁺ memory B cells remained persistently low during the 2 years of follow-up (29). Other authors similarly reported relatively fast naive B cell reconstitution followed by delayed memory B cell recovery (30, 31). In addition, early expansion of CD19⁺CD5⁺ B cells has been reported (29, 32), a subset described as pre-naive circulating B cells representing a distinct intermediate phenotype between transitional and naive B cells (33). These cells showed only partial responses to B cell receptor (BCR) stimulation and CD40 ligation, but similarly to CD19⁺CD21^highCD27⁻ naive B cells, these were capable to differentiate into plasma cells and had the ability to function as antigen-presenting cells (APCs) (33).

In the first 2 years following allogeneic HSCT, B cell function remains compromised. Different B cell subpopulations often reconstitute over a different period of time contributing to a defective humoral response. Delayed T cell recovery and the reversed CD4/CD8 ratio may also contribute to low circulating B cell numbers following HSCT (26). Furthermore, CD19⁺CD27⁺ memory B cells can be influenced by low T helper cells as they require their help for isotype switching (26). In addition, somatic hypermutation seems to be diminished even in the presence of normal donor CD4⁺ T cell numbers, implying an environmental defect (26, 34). Normal levels of serum IgM are usually measurable 3–6 months after HSCT (35, 36), followed by normalization of serum IgG1/IgG3, IgG2/IgG4, and IgA similar to that observed during normal development in the first years of life (37). However, in some patients, long-term antibody class deficiencies have been reported (38). The immunoglobuline heavy chain (IgH) repertoire is often characterized by delayed class switching and oligoclonal dominance with specific rearrangements dominating at different time points in these patients (36, 39). Measurement of B lymphocyte repertoire diversity by analysis of IgH complementarity determining region 3 (CDR3) revealed limited variation of IgH CDR3 repertoire in CD19⁺CD27⁺ memory B lymphocytes compared to CD19⁺CD21^highCD27⁻ naive B cells at 3 and 6 months after allogeneic HSCT. Decrease in CD19⁺CD27⁺ memory B cell IgH CDR3 repertoire, but not CD19⁺CD21^highCD27⁻ naive B cell one, was also observed when compared to healthy controls suggesting a role of CD19⁺CD27⁺ memory B cells in oligoclonal restriction (35). Both CD19⁺CD27⁺ memory B cells and CD19⁺CD21^highCD27⁻ naive B cells reach normal diversity, comparable to healthy individuals, 12 months after HSCT (35).

Different settings of HSCT may also influence B cell recovery. Patients receiving antithymocyte globulin-fresenius (ATG-F) presented delayed CD19⁺ B cell recovery up to 5 months after HSCT compared to non-ATG-F patients (40). ATG is a potent immunosuppressant administrated before HSCT to prevent graft rejection and to reduce incidence of acute and chronic GvHD in patients receiving grafts from unrelated donors (40, 41). Absolute CD19⁺ B cells normalized 1 year after HSCT in both groups. ATG-treated patients had significantly worse CD19⁺CD21^highCD27⁻ naive B cell and CD19⁺CD27⁺ memory B cell regeneration within the first month after HSCT indicating a negative impact of ATG on B cell immune reconstitution (40). Depending on the brand, ATG may also have immunomodulatory effects on B cells (42). Slow B cell recovery has been observed in patients receiving non-myeloablative conditioning compared to those given myeloablative therapy, with reduced B cell numbers observed in most patients up to 12 months after non-myeloablative therapy for HSCT (43). However, these findings may in part be explained by older patient age and higher incidence of acute GvHD in this patient cohort (43). Both acute and chronic GvHD have been associated with delayed B cell reconstitution, and reduction or lack of B cell precursors in the BM has been observed in these patients compared to patients without GvHD (44). In a study on 93 allograft recipients, the number of BM B cell precursors on day 30 after HSCT was significantly lower in patients later developing grades 2–4 acute GvHD compared to patients with grades 0–1 disease (44). Moreover, patients developing extensive chronic GvHD within 1 year after transplantation had lower percentages of B cell precursors on day 365 compared with patients without chronic GvHD or with limited chronic GvHD (44). However, the effect of acute and chronic GvHD could not be separated from the possible influence of glucocorticoid treatment in this study due to low patient numbers suggesting B cell deficiency after transplantation may in part be a result of inhibition of B lymphopoiesis by GvHD and/or its treatment (44). In addition, a decrease of absolute CD19⁺ B cells in patients at first diagnosis of chronic GvHD and a disturbance of B cell homeostasis in patients with active chronic GvHD have been observed (45, 46). Stem cell source may also influence numbers of circulating B cells with higher counts detected in recipients of peripheral blood stem cells (PBSC); however, this observation may be attributed to the higher amount of mature B cells in PBSC grafts compared with BM (44, 47, 48).

Even patients who show recovery of overall CD19⁺ B cell numbers are not considered fully immunocompetent and as a result of decreased B cell function, impaired vaccine responses to infectious antigens have been observed (26). Lack of CD19⁺CD27⁺ memory B cells, decrease of circulating immunoglobulins, and impaired immunoglobulin gene rearrangement render these patients susceptible to encapsulated bacteria and viruses (1, 24).

CD4⁺ and CD8⁺ T cells reconstitute within the first year after HSCT and enable defense against viral or fungal infections, as well as maintaining the GvL effect. A subset of CD4⁺ T cells are so called regulatory T cells (Tregs). In the next paragraph, we aim to summarize their development and function in patients after allogeneic HSCT.

Tregs play an indispensable role in both solid organ transplant tolerance and in allograft tolerance after HSCT. In rodents and humans, a subpopulation of thymus-derived naive CD4⁺ T cells that co-express the IL-2R alpha chain, CD25, have potent suppressive activity (64). Tregs mediate transplantation tolerance in experimental models of skin and/or solid organ transplantation (65) as well as tolerance to BM allografts (66). By allogeneic HSCT, malignant and non-malignant hematological disorders can be cured, but at the same time, treatment efficacy is limited due to occurrence of GvHD (67). Regulatory T cells have received considerable attention in recent years due to their ability to suppress the proliferation of conventional T cells when added to donor grafts and prevention of GvHD in animal models (68). Using a mouse model, Edinger and colleagues have shown that CD4⁺CD25⁺ Tregs suppress GvHD after BMT without abrogating the GvL or graft-versus-tumor (GvT) effect (14) supporting the importance of Tregs in allogenic HSCT. Furthermore, Nguyen and colleagues demonstrated that the adoptive transfer of Tregs preserved thymic and lymphoid architecture of the host and hence accelerated posttransplant T cell immune reconstitution in a murine GvHD model (69).

Taylor and colleagues demonstrated that in vivo depletion of CD25⁺ T cells and depletion of CD25⁺ T cells in the transplant inoculum, worsened GvHD whereas adoptive transfer of CD4⁺ CD25⁺ nTregs together with the BM graft ameliorated GvHD (70).

While an increasing number of publications have focused on the biology of CD4⁺CD25⁺Foxp3⁺CD45RO^lo naive Tregs in GvHD, less attention has been given to iTregs, probably due to the lack of proven cell surface marker that differentiate nTregs from iTregs. Fantini and colleagues demonstrated that iTregs can be generated from CD4⁺ T cells in the presence of TGF-β and can be expanded in culture (71). On the other hand, Koenecke and colleagues showed that administration of in vitro generated iTregs along with BM grafts containing alloreactive donor T cells did not provide any significant protection from lethal GvHD, due to limited in vivo survival of these cells (72). They also demonstrated that iTregs lost their Foxp3 expression, along with a loss of suppressive function early after transplantation, thus making iTregs unsuitable for use in a therapeutic approach (72) if administered as an external cellular product. Not only iTregs but also nTregs have been shown to loose Foxp3 expression in a STAT3-dependent manner and can revert to a proinflammatory phenotype under inflammatory conditions (73). Therefore, inflammation seems to affect Foxp3 expression in both natural and induced Tregs.

Although Tregs have long been assumed to be solely a subset of the CD4⁺ T cell compartment, a CD8⁺ Treg population has been recently described and shown to be capable of suppressing T cell responses (74). In terms of GvHD, Robb and colleagues reported that CD8⁺Foxp3⁺ Tregs suppressed GvHD and attenuated GvHD mortality after BMT in a mouse model (75). Interestingly, CD8⁺Foxp3⁺ cells were more suppressive than CD4⁺Foxp3⁺ cells (75). Using a rat model, Xystrakis and colleagues provided a first report on CD8⁺ Tregs conferring their regulatory properties via a cell to cell contact dependent mechanism to prevent GvHD and thus confirming CD8⁺Foxp3⁺ Tregs in a second species (76). Clinical studies on CD8⁺ Tregs at a functional level are scarce to date. However, Zheng and colleagues reported that human CD8⁺ Tregs potentially inhibit GvHD without compromising general immunity and GvL activity in humanized mouse models (77). Taken together, these findings provide an insight into the efficacy of both CD4⁺ and CD8⁺ Tregs as potential novel therapeutic approaches in clinic.

Many researchers have focused on evaluating Treg cell numbers after HSCT, since they play an important role in the amelioration of GvHD. Using PB of patients after transplantation, Li and colleagues demonstrated that the frequency of CD4⁺CD25⁺ Tregs was significantly downregulated in patients with severe acute or chronic GvHD (78). They also showed that a decreased level of CD4⁺CD25⁺ Tregs correlated with increased severity of GvHD (78). While the majority of studies focused on blood derived Tregs, there is little information on Tregs isolated from intestinal tissues due to the lack of availability of repeated gut biopsies. Using immunoenzymatic labeling, Rieger and colleagues were the first to demonstrate that infiltrating Tregs decreased the signs of acute and chronic GvHD in intestinal mucosa (79). They showed that patients with acute and chronic GvHD had a complete lack of counter regulation indicated by a Foxp3⁺/CD8⁺ T cell ratio identical to that of healthy individuals, while this ratio was increased in patients without GvHD (79). These results have been discussed controversially in the literature since Lord and colleagues demonstrated that Foxp3⁺ T cells were not decreased in PB or gastrointestinal tissues and that the frequency of Tregs did not correspond to disease incidence or severity (80). On the contrary, these investigators reported that Foxp3⁺ T cells were significantly upregulated in GvHD-afflicted intestinal mucosa when compared to non-GvHD mucosal tissues (80). This finding was further supported by Ratajczak and colleagues who observed an increased proportion of CD4⁺Foxp3⁺ T cells in patients with grades 2–4 compared to grades 0–1 acute GvHD (81). One possible explanation for these conflicting results may be the difficulty to discriminate natural and induced Tregs. It is possible that nTregs are decreased in GvHD while iTregs may be increased in order to compensate for the exaggerated inflammation during GvHD. Imanguli and colleagues observed an upregulation of functional markers such as CD3⁺, CD4⁺, CD27⁺, ICOS⁺, and CD39⁺ in Tregs that traffic into tissue including skin and oral mucosa exerting a suppressive function in patients with chronic GvHD (82). Interestingly, normal numbers of activated CD45⁻Foxp3^hi Tregs were observed in tissue and PB of patients with chronic GvHD whereas naive or resting CD45RA⁺Foxp3⁺ Tregs that presumably control chronic GvHD effector cells were reduced compared to patients without chronic GvHD.

Studies in mouse models of GvHD have provided information on the suppressive nature of Tregs and their potential to suppress and ameliorate GvHD without impairing the GvL effect. The first clinical trial using Tregs to suppress acute/chronic GvHD in patients were completed recently. This “first-in-man-study” reported the adoptive transfer of ex vivo expanded CD4⁺CD25⁺CD127⁻ Tregs in one patient with chronic GvHD and another with acute GvHD after HSCT with an HLA-identical sibling donor (83). Transfer of Tregs resulted in a reduction of the steroid dose administered, increased levels of circulating Tregs, and a decrease in inflammatory cytokine levels in the PB (83). Another “first-in-man-study” was reported after double UCBT in 23 patients, who received in vitro expanded 0.1–30 × 10⁵ UCB CD4⁺CD25⁺CD127⁻ Tregs per kilogram derived from partially HLA-matched third-party UCB units (15). There was a significant reduction in the incidence of acute GvHD grades II–IV (43 versus 61%, P = 0.05) when compared to 108 historical controls without transfusion of Tregs. No toxicities, infections, relapse, or early mortality were observed suggesting that UCB Tregs could be beneficial for preventing acute GvHD (15). Furthermore, Di Ianni and colleagues reported a clinical trial in 28 patients receiving adoptively transferred CD4⁺CD25⁺CD127⁻ Tregs after T-cell-depleted haploidentical HSCT without further immunosuppression (13). Only 2 out of 28 patients developed grades II–IV acute GvHD and no chronic GvHD was observed. They showed that adoptive transfer of freshly isolated donor-derived Tregs 4 days before inoculating the CD34⁺ stem cells prevented acute and chronic GvHD in the absence of further immunosuppression. Tregs promoted lymphoid reconstitution, improvement of immunity to opportunistic pathogens (no CMV-related death of patients) without abrogating the GvL effect (13). In addition, Hoffmann reported in vitro expansion of highly purified polyclonal human CD4⁺CD25^high Tregs through the use of artificial APCs for repeated stimulation via CD3 and CD28 in the presence of high-dose IL-2 (84). These cells not only maintained their phenotype and expressed suppressive activity but also maintained the expression of the lymph node homing receptors l-selectin and CCR7 (84). Furthermore, the same group reported results of a small phase I safety and feasibility trial where freshly isolated donor-derived CD4⁺CD25^high Tregs were infused into nine patients with high risk for leukemic relapse after cessation of systemic GvHD prophylaxis (12). After 8 weeks, additional CD4⁺ T cells were administered to promote GvL activity. Patients showed no signs of GvHD nor opportunistic infections or early disease relapse supporting the safety and efficacy of Treg transfusion (12). This has led to a phase II clinical trial for the treatment of patients with steroid-refractory acute GvHD using freshly isolated CD4⁺CD25^high Tregs that is currently ongoing. Taken together, these early trials suggest that Tregs could be a novel approach for prophylaxis and treatment of patients with acute GvHD in larger clinical trials. The impact of Treg transfusion on the immune reconstitution has to be further investigated.

Tregs induce tolerance and maintain immune homeostasis (51). A major challenge of Treg cell therapy is their relative scarcity in PB (0.5–1% of CD4⁺CD25^high T cells) (85). In 2011, Hippen and colleagues presented two individual reports regarding the generation of induced Tregs on a large scale (86) and ex vivo expansion of natural Tregs (86). Both methods focus on the development of expansion protocols for either type of Tregs to obtain higher yields for clinical trials on treatment or prevention of GvHD (86). In patients with chronic GvHD, Matsuoka and co-workers reported that daily administration of low-dose IL-2 induced selective expansion of functional CD4⁺CD25⁺CD127⁻ Tregs, improved chronic GvHD, restored CD4⁺ T cell homeostasis, and promoted the reestablishment of immune tolerance (87). Koreth et al. reported the case of 29 chronic GvHD patients that the administration of subcutaneous low dose IL-2 rapidly induced preferential and sustained Tregs expansion without any immune impairment (88). This suggests that low-dose IL-2 could be a potential therapy to restore immune balance after HSCT. Another approach to manipulate Tregs in vivo was reported by Furusawa and colleagues (89). Clostridial products, like short chain fatty acids (SCFA) or mainly butyrate, can induce the differentiation of colonic Tregs in vitro and in vivo in mouse models (89). This points toward the necessity of host–microbiome interaction to establish immunological tolerance and homeostasis in the gut. Moreover, Mathewson and colleagues reported that restoring clostridial metabolites or the strain itself modulated intestinal epithelial cell integrity and mitigated GvHD in mice (63). Taken together, these findings strongly suggest that the right balance of gut microbiome may be crucial to induce Tregs for intestinal tolerance.

Memory T cells [central memory (T_CM) and effector memory (T_EM)], tissue resident memory cells (T_RM), and effector cells (T_EFF) cells are essential to control viral reactivations after allogeneic HSCT. Upon encountering antigens, memory cells differentiate to T_EFF and lyse the infected cells and secret proinflammatory cytokines [e.g., IFN-γ and tumor necrosis factor-α (TNF-α)] (90). Immune surveillance of T_CM occurs trafficking through secondary lymphoid organs, T_EM and T_EFF, through non-lymphoid organs (91). In contrary, T_RM cells reside at various sites (e.g., liver, lungs, gut, and skin) and provide immediate antiviral response (cytotoxicity and secretion of IFN-γ) without trafficking (92). The reconstitution of CD4⁺CD45RA⁺ naive T cells, providing the broad range of TCR repertoire needed to control infections and to avoid the reappearance of leukemic cells, is essential after allogeneic HSCT (11, 93). The conditioning regimens applied, increasing patient age and occurrence of acute and chronic GvHD, have devastating effects on thymic function after HSCT (28, 93–95). Reconstitution of CD4⁺CD45RA⁺ naive T cells can be demonstrated by measuring TRECs. Immune reconstitution of CD4⁺ and CD8⁺ T cells is also essential for maintaining a GvL effect (1). Reconstitution of CD8⁺ T cells is faster than that of CD4⁺ T cells, which usually occurs around day +100 or later and is indicated by the inversion of the CD4⁺/CD8⁺ T cell ratio (1) early after HSCT (Table 1). The time period until complete reconstitution of CD4⁺ T cells can take up to 2 years after allogenic HSCT (96).

Graft-versus-leukemia is defined as an immune response directed against leukemia/tumor cells of the recipient after allogeneic HSCT. Over the years, several studies have shown that CD4⁺ and CD8⁺ T cells play an important role in establishing a GvL effect through various mechanisms such as cytotoxic T cells releasing granzyme B and apoptosis mediated by FAS ligands (115). GvL is often associated with GvHD, but GvL responses against, e.g., minor histocompatibility antigens solely expressed on hematopoietic cells (mHA1) may be specifically directed against leukemic cells or the recipients’ hematopoietic cells. The precise role of CD4⁺ and CD8⁺ T cells for achievement of a GvL effect is not clearly understood today (115, 116). Complete depletion of T cells by CD34⁺ cell selection leads to a high incidence of relapse, resulting in death in about 20–50% of patients (117). T cell repletion or donor lymphocyte infusions (DLI) can prevent relapse, but can lead to a higher probability of acute and chronic GvHD (118, 119). Several protocols tried to circumvent the problem of increased acute and chronic GvHD by delayed add-back of genetically modified T cells (109) or other manipulations of the donor’s lymphocytes such as selection of either CD4⁺ or CD8⁺ T cells prior to transfusion (120, 121).

T cells are the most important effector cells in the control of viral infections. Thus, T cell reconstitution after allogeneic HSCT has a significant impact on the control of infectious complications. The first phase of virus-specific T cell reconstitution and expansion early after HSCT depends on the transfer of mature (effector, memory, or naive) virus-specific T cells within the donor graft and the resident antigen-specific cells (10, 122). Viral infections occur mostly between engraftment and day +90 posttransplant (123). However, also late (after day +90) and recurrent CMV reactivations have been observed, which have been associated with impaired reconstitution or function of antiviral immunity (106). CMV is a latent virus, which belongs to the family of herpesviruses and is one among the common viral pathogens that can reactivate after HSCT. It reactivates in about 60–70% of CMV-seropositive patients, and the primary infection affects 20–30% of CMV seronegative recipients transplanted from CMV-seropositive donors (124). Uncontrolled CMV reactivations can lead to a life-threatening, multi-organ CMV disease such as retinitis, gastroenteritis, or pneumonia (125–127). Advances in CMV monitoring, preemptive antiviral therapy, and quantification of CMV–CTLs are crucial in the prevention of CMV disease (128). The most important risk factors for CMV infection include recipient CMV-seropositivity, TCD of the graft, and acute GvHD (123). Early reconstitution of antiviral immunity remains an essential issue for the control of CMV reactivations after HSCT. The recovery of both CD8⁺ and CD4⁺ CMV-specific T cells may be a marker for protection against CMV reactivations (129).

Epstein–Barr virus infection is also a frequent viral complication after allogeneic HSCT, which may progress to EBV-associated posttransplant lymphoproliferative disease (PTLD) that causes unspecific symptoms such as fever and lymphadenopathy with a high viral load in the PB (130). These complications are mediated by several risk factors including TCD combined with reduced intensity conditioning (RIC) leading to impaired anti-EBV T cell-mediated immunity and persistence of residual recipient B cells (131). In addition, HLA disparity and acute GvHD have also been known to increase the risk of PTLD due to the delayed or impaired specific immune reconstitution (132).

However, other viral pathogens such as adenovirus (ADV), human herpes virus 6 (HHV6), BK-polyoma virus (BKV), and respiratory viruses occur less frequently in adult patients in comparison to CMV and EBV after allogeneic HSCT (133). The control of these viruses again depends upon the reconstitution of antiviral immunity.

Among the viruses mentioned above, T cell immune reconstitution against CMV has been studied most intensively and will be described in more detail below, as an example for virus (or any antigen)-specific T cell reconstitution and expansion. Apart from the above mentioned factors influencing T cell reconstitution (TCD of the graft; stem cell source, occurrence of acute or chronic GvHD), CMV serostatus of patient and donor is one of the most important variables influencing CMV-specific T cell immune reconstitution. CMV-seropositive recipients and donors (R+D+) have much faster reconstitution of CMV–CTLs (prior to day +50) and a subclinical CMV reactivation can even boost this development (106). On the other hand, CMV-seropositive recipients transplanted from CMV-seronegative donors (R+D−) lack the protective donor-derived immunity and hence have delayed recovery of antiviral immunity (between days +120 and +150) and a higher risk for recurrent CMV reactivations (134). In Figure 4, examples of patients with typical CMV–CTL immune reconstitution for R+D+ and R+D− groups are shown to demonstrate the impact of CMV serostatus on CMV–CTL immune reconstitution. Additionally, it has been shown, that CMV–CTLs of recipient origin can survive the conditioning regimen and can add to the protection against CMV, especially in R+D− patients (135).