Allogeneic haematopoietic stem cell transplantation (HSCT) is a standard form of treatment for patients diagnosed with haematological disorders, including malignancies. Even though it is a common procedure, it may lead to development of serious complications, such as graft-versus-host disease (GvHD), which may occur in an acute (aGvHD) or chronic (cGvHD) form. Acute GvHD is initiated as a reaction of donor cells towards tissues of the recipient. It affects skin, gut, lung and liver depending on the severity of the disease, which is graded I to IV. Chronic GvHD has a different pathogenesis and can affect more organs than aGvHD (1, 2). Aside from GvHD, another type of post-transplant complications is viral infections, particularly those caused by Herpesviridae such as cytomegalovirus (CMV) or Epstein-Barr virus (EBV). Asymptomatic, latent CMV infections are extremely common, with as many as 50% of people being seropositive for CMV (3). However, CMV reactivation may be life-threatening in immunocompromised persons, such as post-HSCT patients (4).

Natural killer (NK) cells, as the first donor-delivered lymphocytes to reconstitute after HSCT, have protective properties against GvHD with a simultaneous ability to induce a graft-versus-leukaemia (GvL) effect (5–7). NK cells are orchestrated by a wide set of activating and inhibitory receptors, whose ligands are classical and non-classical Major Histocompatibility Complex (MHC) molecules, e.g. HLA-E. In contrast to other MHC molecules, HLA-E is very conserved and its polymorphism is mostly limited to two major alleles, *01:01 and *01:03, comprising over 99% of allele frequency globally (8). Both of these alleles are distributed with similar frequencies and differ in a single Arg/Gly substitution in position 107. Other HLA-E alleles exist, although they are extremely rare. Some of these alleles also characterize with alternative substitutions in position 107 (e.g. HLA-E*01:48, with allele frequency of 0.0007%) (9). HLA-E serves as a ligand for two NKG2 receptors, inhibitory NKG2A and activating NKG2C (encoded by the KLRC1 and KLR2C genes, respectively). It can also be secreted in a soluble form (sHLA-E). This soluble form may play a role in immune regulation (10), and sHLA-E levels are increased in various cancers and autoimmune diseases (11–14). HLA-E:NKG2A/C interactions are essential for balancing the NK cell reactivity (15). Despite the molecular and structural similarities of these two receptors, HLA-E binds NKG2A with 6-fold higher affinity, what helps to monitor the expression of the MHC class I molecules on normal cells (16). NKG2C expression is low in immature NK cells, and then subsequently increases during maturation, while NKG2A expression concurrently decreases (17). A specific subset of NKG2C+ cells has been observed to expand in response to cytomegalovirus (CMV) reactivation, but not other viral infections such as EBV. These cells can function like adaptive memory cells, and, if transplanted from CMV seropositive donors, exhibit a heightened response to a secondary CMV event (18–21). They persist for a long time after infection and lack NKG2A expression (20). NKG2C+ NK cells were also observed to interact with CMV-specific CD8+ T cells to combat CMV infection (22). Both NKG2A and NKG2C are minimally polymorphic compared to classical HLA genes. There are several rarely studied single nucleotide polymorphisms (SNPs) in the NKG2C gene, some of which are located in coding regions, associated with three alleles designated as NKG2C*01, NKG2C*02 and NKG2C*03 (23, 24). In contrast, a major NKG2C deletion resulting in a loss of expression or reduced expression (in homo- and heterozygotes, respectively) has been far better studied and is well known for its importance in viral infections (25). As reported in many studies (26–28), patients carrying at least one del variant are more susceptible for CMV (especially reactivation after HSCT), and HIV infections, nonetheless the studies are not always consisted (29). It has been recently suggested that the NKG2C deletion may also increase risk for SARS-CoV-2 infections, although this seems to require further validation (30).

The selected HLA-E rs1264457 SNP is localized in the third exon. It results in a T (*01:01)/C (*01:03) nucleotide substitution associated with Gly to Arg amino acid exchange. The selected polymorphism for the NKG2A inhibiting receptor was rs7301582, an intronic C/T substitution. Both SNPs (HLA-E rs1264457 and NKG2A rs7301582) have also been studied in our recent study on post-transplant complications in paediatric HSCT recipients (31), while in our previous work on patients with inflammatory (rheumatoid) arthritis, they were described to be associated with response to anti-TNF treatment (32, 33). Another authors investigated these two SNPs in their studies, as e.g. Kordelas et al., who proved a protective effect of HLA-E*01:03 homozygosity in overall survival after HSCT (34, 35).

We hypothesize that HLA-E, NKG2A and NKG2C expression and polymorphism play a role in the development of complications after HSCT in a Polish population. Various studies previously described the role of HLA-E and NKG2A/C polymorphisms in different population groups, although their results were not always consistent (28, 36–44). In our present study, we aimed to determine the role of HLA-E genetic polymorphism and soluble HLA-E concentration, as well as NKG2A and NKG2C gene polymorphisms and protein expression within the NK cells in the development of post-transplant complications in recipients of allogeneic hematopoietic stem cells.

The study group consisted of 200 HSCT recipients (aged 19-73) and 104 of their donors treated in five Polish transplantation centres. Patients were assigned for allogeneic HSCT according to European Society for Blood and Marrow Transplantation criteria. Exclusion criteria were: age < 18 years old, high Haematopoietic Cell Transplantation-specific Comorbidity Index (HCT CI), and Karnofsky index < 80%. The most common haematological disease was acute myeloid leukaemia (AML), diagnosed in 42.5%, followed by acute lymphoblastic leukaemia (ALL), diagnosed in 13.5% of recipients. The patients characteristics and transplant details are presented in Table 1 . Recipients and donors were genotyped at high resolution level for at least 5 HLA loci (HLA-A, B, C, DRB1 and DQB1). This study complies with the Declaration of Helsinki and was approved by the Wroclaw Medical University Ethics Committee (identification code KB-561/2019).

Since GvHD and viral infections remain the most common HSCT complications, the need to develop new forms of treatment and characterize new prognostic as well as risk factors is a high priority. The HLA-E: NKG2A/C interaction may be a key element of these processes.

A number of studies report that donor and recipient HLA-E*01:03 homozygous genotype is related with better HSCT outcome due to a lower risk of acute and chronic GvHD development and a lower rate of relapse (48–53). However in contrast, in T-cell-replete transplantations, donor HLA-E*01:01/*01:03 genotype was associated with a higher transplant-related mortality and lower disease-free survival (36). Also, a study on kidney transplant patients showed that HLA-E*01:03 allele was correlated with a higher incidence of CMV infection after transplantation (37). A similar association was observed between HLA-E*01:03 and hepatitis C viral infection (38). This shows that results of previous studies on HLA-E polymorphism are somewhat inconsistent. Interestingly, our present study did not show any association between the presence of either HLA-E*01:03 or HLA-E*01:01 allele and a risk of post-transplant complications. However, we observed that HLA-E mismatch between donor and recipient is associated with a higher risk of CMV infection after HSCT. Likewise, an earlier study observed that HLA-E mismatch affected HSCT survival in patients with acute leukaemias, particularly in patients with advanced disease (54). In our previous studies, we also observed the unfavourable effect of HLA-E mismatching in patients after HSCT, such as increased acute GvHD risk (55–57). This suggests that HLA-E incompatibility, in addition to the presence or absence of specific alleles, may be an important factor affecting HSCT outcome.

Production of cell-free soluble HLA molecules is known to be implicated in cancer immune escape (58). HLA-E can likewise be present in a soluble form, resulting largely from proteolytic cleavage of membrane-bound HLA-E by matrix metalloproteinases (59). Increased soluble HLA-E concentration was associated with disease susceptibility in various disorders, including chronic hepatitis B, juvenile idiopathic arthritis, and gastric cancer (60–62). Soluble HLA-E is also involved in endothelial cell activation and may have a role in immunoregulatory functions of the endothelium (10). Furthermore, increased sHLA-E was detected in various different tumour cell culture supernatants and was proved to be upregulated by various cytokines, such as IFN-γ, IFN-α and TNF-α (11). Conversely, we showed that recipients with GvHD characterized with decreased serum sHLA-E levels 30 days after transplantation. This association was especially seen in patients with chronic GvHD, and to a lesser extent in patients with severe acute GvHD. Similarly, a recent study by Kordelas et al. showed that HSCT recipients who developed acute or chronic GvHD had lower sHLA-E levels up to one year after transplantation (35). These results suggest that decreased sHLA-E serum concentration may serve as prognostic factor for the development of GvHD with lower levels being associated rather with unfavourable prognosis. We also noticed that sHLA-E concentrations increased over time after HSCT, irrespective of the presence or absence of post-transplant complications. This observation might suggest that higher sHLA-E concentrations later on after transplantation may be related with better transplant outcome. Obviously this hypothesis requires further and more comprehensive studies.

Expression of the activating NKG2C NK cell receptor was shown to decrease a year after HSCT, following development of both acute and chronic GvHD in previous studies (39–41). This was not observed in our study, although it may be possible that this discrepancy was due to a shorter time of observation (3 months) of our patients. Conversely, NKG2C expression on NK cells is known to increase during CMV infection (19, 63), leading to emergence of potent mature NKG2A-/NKG2C+ CD56dim NK cells a year after transplantation (64). This is in accordance with our results showing a notable increase in NKG2C expression on all NK cell subtypes at the third month after transplantation as well as an increase of the unique NKG2A-/NKG2C+ NK cell population. The presence of the NKG2C deletion was previously reported to affect CMV reactivation after haploidentical HSCT or lung transplantation (28, 42, 43). We did not observe such an association. However, our study revealed that in patients with post-transplant CMV infection who carried at least one NKG2C del allele, the frequency of NKG2C+ NK cells was decreased in comparison to wild type homozygotes. This observation proves a functional association between the NKG2C gene polymorphism and its expression in patients with CMV infection. Interestingly, this finding seems to be confirmed in part by an earlier study on NKG2C deletion in CMV seropositive children (65). Although this association between NKG2C deletion and decreased NKG2C+ NK may be expected, there are studies indicating that it’s not always present (44).

Little is known about the NKG2A rs7301582 polymorphism. The C variant was reported to have a negative impact on response to anti-TNF treatment in Polish patients with rheumatoid arthritis (33). The present study is, to the best of our knowledge, the first to indicate the potential significance of NKG2A rs7301582 polymorphism in HSCT. It is uncertain how this SNP could exert its effect, since it is located in an NKG2A gene intron (intron 6). It is possible that this variant could affect splicing, although there is no evidence for it. Introns are known to affect gene expression by various indirect means, e.g. changing mRNA stability, influencing methylation/chromatin modifications, or harbouring cryptic splice sites (66, 67). Furthermore, SNPs can also act affect expression of other remote genes in a trans manner (68). Here, we confirm the role of NKG2A rs7301582 C allele as a negative factor since it occurred more frequently in recipients diagnosed with AML than in donor group.

Our study has, however, some limitations, i.e. small sample size, non-homogenous recipient group and lack of availability of all donor samples. Being aware of these limitations, we still are assured that our results make a strong contribution to the development of the field.

Taken together, the results obtained in this present study imply that the donor/recipient HLA-E mismatch is associated with a higher incidence of CMV infection after transplantation. Decreased serum sHLA-E concentration is associated with development of both chronic and acute GvHD in adult HSCT recipients. We also observed significant differences in the frequency of NKG2C expressing NK cells in the context of CMV infection. Recipients who were diagnosed with this complication were characterized by increased percentage of NKG2C+ NK cells. Moreover, we proved that NKG2C deletion was associated with expression level of the receptor, as patients carrying the del allele had a decreased frequency of NKG2C+ NK cells. Regarding the NKG2A polymorphism, the rs7301582 C allele may be associated with higher AML susceptibility. In conclusion, our results suggest that the sHLA-E level and expression of NKG2C on NK cells may act as a potential markers of post-transplant complications.

The datasets presented in this study can be found in online repositories. The data can be found at: https://cloud.hirszfeld.pl/index.php/s/sFaqtFN4wwiKWbK. Further inquiries can be directed to the corresponding author.

The studies involving humans were approved by Wroclaw Medical University Ethics Committee. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

JS performed genotyping studies and assessment of sHLA-E concentration in serum samples, statistical analyses, drafted and finalized the manuscript; PŁ contributed to statistical analyses; DS performed flow cytometry experiments; AC, AS, MM, MS-K, WF, IS, BN-A, PS, MB, AT, GB, SG, TW provided patients’ clinical samples and clinical data; AC contributed to the conception of clinical data analysis; KB-K contributed to the conception and design of the study, data analysis, drafted and finalized the manuscript, and provided funding. All authors approved the final version of the manuscript.