Allogeneic hematopoietic stem cell transplantation (allo-HSCT) has emerged as an effective treatment for refractory or relapsed hematological malignancies, including leukemia, lymphoma and myelodysplastic syndrome (1). Despite its therapeutic benefits, graft-versus-host disease (GVHD) remains one of the most significant complications following HSCT and a leading cause of transplant-related morbidity and mortality. The prognosis for patients with severe (grade III-IV) GVHD remains particularly poor, with reported survival rates varying from 25% in the adult population to 55% in pediatric cohorts (2). While the treatment with immunosuppressants has reduced the incidence and mortality of GVHD, this has been accompanied by an increased risk of other clinically significant complications, including opportunistic infections (particularly cytomegalovirus (CMV) and Epstein-Barr virus (EBV) reactivation), hemorrhagic cystitis (HC), and capillary leak syndrome (CLS) (3).

Calcineurin inhibitors (CNIs), like cyclosporine A (CsA), remain a cornerstone in both the prophylaxis and treat GVHD after HSCT, which significantly enhanced the overall survival (OS) rates after transplantation (4). Cyclosporine A (CsA) is an 11-amino acid cyclic peptide obtained from fungal fermentation. CsA works by selectively inhibiting the activity of calcineurin, which suppresses the lymphocyte activation and proliferation, leading to its immunosuppressive effects (5). The early post-transplant concentration of CsA is closely associated with GVHD risk, highlighting the importance of individualized drug management (6).

CsA is mainly metabolized by enzymes from the CYP450 family, predominantly CYP3A4 and CYP3A5 (7). For example, CYP3A5*3C, a splicing variant in intron 3, leads to aberrant mRNA splicing and a premature stop codon, resulting in a truncated, non-functional protein (8). Several studies have demonstrated that the CYP3A5*3C/*3C genotype correlates with a higher CsA C₀/dose in early renal transplant patients (9, 10). Moreover, CYP3A4*1G and CYP3A5*3C are in strong linkage disequilibrium in Asians, and individuals carrying the CYP3A5*1 allele are more likely to co-inherit the CYP3A4*1G allele than the CYP3A4*1 allele. These indicate differences in individual single nucleotide polymorphism (SNP) are the primary factors affecting the metabolic activities of these enzymes.

In addition to metabolic enzymes, CsA acts as a substrate for various influx and efflux transporters, such as ABCB1 and SLCO1B1 (11, 12). SNP variations in transporter coding gene, such as ABCB1, have been associated with altered CsA pharmacokinetics. Notably, ABCB1 3435CC genotype carriers tend to require higher CsA doses than 3435TT carriers, particularly in Asian populations and during the early post-transplant period (13). Other metabolic pathways, including UGT1A8 and BMP-related pathways, may also contribute to CsA disposition, although these associations remain less well defined (14, 15).

However, limited reports have examined the relationship between gene SNPs and clinical complications after transplantation.Our study aimed to identity gene SNPs of CsA-related transporters and metabolic enzymes, including ABC, BMP, CYP, MTHFR, MTRR, SLC, STIM1, UGT1A8 family, and correlate these SNPs with common post-transplant clinical complications in 128 pediatric hematological malignancy patients.

Various SNPs are associated with changes in the drug metabolism and effects, explaining differences in drug responses among individuals (28, 29). SNPs affecting the metabolism of drugs like CsA, used in GVHD prophylaxis, have been related to post-HSCT complications (26).

Our study investigates the correlation between CsA-related SNPs and the transplantation outcomes in 128 patients with malignant hematological diseases patients who underwent HSCT. The incidences of various complications were as follows: grades II-IV aGVHD occurred in 47.7% of cases, grades III-IV aGVHD in 14.1%, cGVHD in 47.7%, CMV infection in 57.8%, EBV infection in 53.1%, HC in 29.7%, and CLS in 16.4%. These complication rates are consistent with previously reported data in the literature (21, 30, 31).

The multivariate analysis in this study identifies UCB-HSCT as an independent factor for peri-ES, consistent with previous reports. This observation may due to the critical role played by granulocyte-macrophage colony-stimulating factor, produced by cord blood-derived inflammatory monocytes, in driving the pathology of peri-ES in UCB transplantation (32). Also, the donor-recipient HLA matches of 9/10-10/10 have been identified as an independent protective factor for peri-ES. Previous studies indicate that mismatched donor-recipient HLA matches in allo-HSCT result in poor transplant outcomes, high incidence and severity of GVHD, slow immune reconstitution, and increased risk of severe infections (33). Some studies also suggest that peri-ES represents as an early form of aGVHD, both being immune reactions occurring post-transplant, implying a certain correlation between the occurrence of peri-ES and HLA matching (34, 35). We observed a significant decrease in CLS incidence with donor-recipient HLA matches of 9/10-10/10. Research indicates that calcineurin inhibitors, like cyclosporine and tacrolimus, used during transplantation and donor graft implantation, can activate and damage endothelial cells (36, 37). HLA matches of 9/10-10/10 can effectively reduce immunosuppressant use, decrease the GVHD incidence, endothelial cell damage and CLS incidence. Moreover, Peri-ES, CLS, and GVHD are increasingly recognized as complications driven by the release of inflammatory cytokines (38–40). It is well established that a higher degree of HLA matching results in reduced activation of donor T cells. In the early post-transplant period, donor T cells secrete pro-inflammatory cytokines such as TNF-α, IL-1, and IFN-γ, which contribute to tissue inflammation and endothelial activation (41). HLA mismatch exacerbates this cytokine storm, leading to endothelial injury. Future prospective studies are warranted to investigate the cytokine profiles and immune cell subset dynamics associated with these complications.

Regarding GVHD, the study identifies several independent risk and protective factors. This study is the first to identify independent risk factors for grades II-IV aGVHD at the time of transplantation, including recipient weight ≤ 26 kg, major and minor ABO blood type mismatched between donor and recipient, and CYP2C19 (99T>C) TC/CC genotype. The relationship between recipient weight and GVHD remains controversial. Mehdi Yaseri et al. conducted a meta-analysis examining the relationship between obesity and post-transplant mortality and clinical outcomes in children, finding an increased risk of post-transplant mortality and aGVHD in obese children (42). Additionally, Lam T Khuat et al. suggested that obesity may increase the risk of aGVHD by affecting the microbiota composition (43). However, other studies argue that obesity has no impact on aGVHD (44). Currently, no studies have shown a correlation between low weight and high incidence of aGVHD. However, recipients with lower weight may still face GVHD-related issues, including immune reactions caused by the immunological imbalance between the graft and the host. Lower weights may result in an insufficient immune system to regulate these reactions effectively. Nevertheless, research in this area is relatively scarce, and further in-depth studies are needed to confirm this relationship. This study identified a significant increase in the incidence of grades II-IV aGVHD with major and minor ABO blood type mismatched between donor and recipient, which is identified as an independent risk factor, which aligns with previous research (45). ABO antigens are expressed on the surfaces of tissues like the gastrointestinal tract, liver, and skin, and they retain their original types due to the newly implanted hematopoietic system. Therefore, donor T lymphocytes will attack the incompatible antigens on tissue surfaces, leading to transplant rejection of corresponding organs. Hence, theoretically, HSCT patients with ABO blood type mismatched have a higher risk of GVHD occurrence. CYP2C19 enzyme is highly polymorphic, and its encoding gene is located on human chromosome 10. CYP2C19 SNPs cause variations in enzyme activity, which affects the concentration of related drugs in the body, thereby affecting their efficacy and may lead to serious adverse reactions (46). The CYP2C19 (99T>C) variants, known as CYP2C19*2, reduce the metabolic rate of drugs, including voriconazole and proton pump inhibitors. Following HSCT, cyclosporine is often co-administered with voriconazole or proton pump inhibitors, which can alter its concentration (47). Therefore, we speculate that CYP2C19 (99T>C) variants may influence cyclosporine concentrations through drug interactions, thereby affecting aGVHD, which is an unexpected finding requiring further exploration and validation in subsequent studies.

Recent studies have identified that risk factors for EBV infection in patients after allo-HSCT, including the use of ATG, selective depletion of T cells, HLA mismatched, age ≥ 50 years old at the time of transplantation, and secondary transplantation (48). Since most patients lack fully HLA matched sibling donors, the main obstacles to using such donors for HSCT are graft rejection and GVHD (49–51). Strengthening pre-transplant conditioning regimens and T-cell depletion transplantation have reduced the incidence of these complications but may lead to slow immune reconstitution and increased rates of various infections. Therefore, we hypothesize that HLA mismatching increases the risk of EBV infection. In this study, we observed a significant reduction in EBV infection rates among recipients with HLA matches of 9/10-10/10. Similar to EBV infection, CMV reactivation is closely associated with the use of immunosuppressants. A previous study indicates that, in kidney transplant patients receiving immunosuppressive regimens with CsA, those with CYP3A5 (219-237C>T) CC genotype exhibited higher dose-adjusted trough concentrations and dose-adjusted peak concentrations of CsA than those with variant genotypes. Consequently, patients with CYP3A5 (219-237C>T) CC genotype may experience greater fluctuations in CsA dose adjustment to achieve target concentrations, thereby increasing the risk of CMV infection (25).

HC is a common complication of HSCT, with previous studies identifying factors as associated with its occurrence, including type of transplantation, age at transplantation, presence of GVHD, donor source, and the composition and intensity of the pre-transplant conditioning regimen (52–54). Consistent with prior literature, our study showed that male recipients and older children at the time of transplantation have a significantly higher incidence of HC (55). This could be attributed to estrogen’s protective roles in stabilizing microvasculature and the bladder mucosa, thereby increases the risk of HC in males. Moreover, older children experience higher rates of HC compared to younger children and are more likely to develop severe HC. This may be related to the less mature urinary and central nervous systems in younger children, which lead to more frequent urination, reducing the dwell time of drugs in the bladder and thereby minimizing the irritation and damage to the bladder mucosa by toxic metabolites, ultimately lowering the incidence of HC. Transporters such as the ATP-binding cassette family and the solute carrier family play key roles in drug transport and absorption. Additionally, this study identified wild-type ABCB1 (1236C>T) and SLCO1B1 (1865 + 4846T>C) as independent risk factors for developing HC.

Previous studies have found that CsA inhibits various influx and efflux transporters, including ABCB1 and SLCO1B1, thereby increasing the area under the curve of other substrates such as bosentan, carprofen, and methotrexate. The degree of inhibition varies with different transporter polymorphisms, leading to variable increases in substrate concentrations (56–58). It was demonstrated that the increase in repaglinide concentration was 42% lower in subjects with the SLCO1B1 (521T>C) variant genotype compared to those with the SLCO1B1 (521T>C) TT genotype. This observation is due to the reduced function of OATP1B1 in individuals carrying the variant SLCO1B1 c.521C allele (59). Therefore, we hypothesize that the wild-type ABCB1 (1236C>T) and SLCO1B1 (1865 + 4846T>C) alleles may exhibit higher activity than variant alleles. This could result in CsA significantly increasing the concentrations of other ABCB1 and SLCO1B1 substrates, including immunosuppressants (MTX, tacrolimus), leading to BK virus activation and subsequent HC. On the other hand, HC is influenced by various factors, often associated with high-dose pre-transplant chemotherapy toxicity, genetic polymorphisms in drug-metabolizing enzymes, viral infections, GVHD, patient age, gender, donor type, and transplantation method. Future studies should incorporate additional factors and expand sample sizes for further validation.

Furthermore, the present study identified that the SLC29A1 (-162 + 228A>C) AC/CC genotype is an independent protective factor for both DFS and OS. SLC29A1 encodes human equilibrative nucleoside transporter 1 (hENT1) (60). As an equilibrative nucleoside transporter, hENT1 facilitates the influx of approximately 80% of cytarabine (Ara-C) into leukemia cells. A study involving 103 AML patients aged 16 to 76 years revealed that SLC29A1 variants and haplotypes may influence the Ara-C uptake activity and the complete remission rate (60, 61). Wan et al. analyzed 19 SLC29A1 SNPs in AML patients and found patients with low SLC29A1 gene expression had shorter DFS and OS during Ara-C treatment. Moreover, significant differences were observed in the expression of the rs9394992 and rs324148 genotypes of SLC29A1 gene between remission and relapse phases (62). The SLC29A1 (-162 + 228A>C) variant, also known as rs693955, has been reported to be associated with a shorter duration of neutropenia following chemotherapy (63). However, no studies have yet demonstrated the relationship between this locus and prognosis. We hypothesize that the SLC29A1 (-162 + 228A>C) variants may influence DFS and OS by affecting the hENT1 protein, thereby interfering with the absorption of cytarabine. In the future, we may verify this hypothesis by measuring plasma cytarabine levels using high-performance liquid chromatography.

However, our study has several limitations. First, the sample size is relatively small. Second, potential confounding factors, such as donor source variability, prior treatments, and differences in post-transplant care, could not be fully accounted for. Third, due to the retrospective nature of the study, external validation in independent cohorts and functional experimental confirmation are currently lacking. To address these issues, we will conduct a prospective, multicenter study with expanded sample size to validate our findings, construct clinically applicable predictive models, and establish mouse models of bone marrow transplantation harboring specific SNPs to investigate the pharmacokinetics and pharmacodynamics of CsA in the future.

In conclusion, our study has revealed correlations between CsA-related transporters and metabolic enzymes SNPs and post-transplant complications and prognosis, contributing to a better understanding of the interindividual difference in drug efficacy. Future studies on adjusting the dosage of drugs based on SNPs in clinical practice may be one of the options for improving the HSCT outcomes.