Introduction

Natural killer (NK) cells, as a type of lymphocyte, might act early in the immune response against tumor cells, and the absolute number of NK cells approximately 4 weeks after transplantation might exceed the number found in healthy individuals (Terszowski et al. 2012). Ruggeri et al. (2002) proposed that donor-derived NK cells can potentially show alloreactivity that results in graft-versus-leukemia (GVL) effects without causing graft-versus-host disease (GVHD). The NK cell receptor repertoire is individual and mainly determined by the KIR genotype. Therefore, the function of NK cells is regulated by inhibitory and activating receptors for major histocompatibility complex (MHC) class I molecules (Farag et al. 2002). KIR gene clusters are divided into haplotypes A and B. Haplotype A mainly encodes iKIRs, with the exception of activating KIR (aKIR) KIR2DS4, whereas haplotype B displays greater variability and includes KIRs that are not present in haplotype A, such as KIR2DL2, 2DL5, 2DS1, 2DS2, 2DS3, 2DS5, and 3DS1 (Hsu et al. 2002).

We previously studied the distribution of specific donor-derived KIR genotypes after unrelated allogeneic hematopoietic stem cell transplantation (allo-HSCT) (Zhang et al. 2017). In recent years, the use of transplant donors possessing KIR haplotype B has been associated with a better long-term prognosis than the use of donors with KIR haplotype A (Bao et al. 2015; Cooley et al. 2014; Symons et al. 2010; Neuchel et al. 2017; Mehta et al. 2018). Elfishawi et al. (2017) reported that KIR2DS3 derived from donors can encourage rapid leukocyte engraftment. Using matched sibling donors, Stringaris et al. (2010) found that donor KIR2DS1 and KIR3DS1 are associated with reduced relapse in acute myeloid leukemia (AML) patients. We previously reported that donor KIR centromeric B-specific gene motifs exert a positive effect on the survival of standard risk AML/myelodysplastic syndrome (MDS) patients (Bao et al. 2015). Haplotype B includes more aKIR genes than haplotype A, and thus provides a basis for the expanding body of research on aKIRs unique to the donor and the corresponding prognoses. Hence, we separately quantified the mRNA levels of iKIRs and aKIRs by RT-qPCR and established reference intervals for each functional KIR gene in the donors and recipients at different timepoints. Our aim was to observe and compare the changes in the mRNA expression of KIR genes, particularly those found only in the donor, and evaluate their utility as prognostic markers after transplantation.

Materials and methods

Patients, donors, and clinical characteristics

Samples were collected from 252 donor-recipient pairs undergoing allo-HSCT between August 2011 and December 2018 at the First Affiliated Hospital of Soochow University (Table 1). The 252 enrolled patients consisted of 161 males and 91 females with a median age of 31 years (range, 8~63 years). The donors, who consisted of 174 males and 78 females with a median age of 31 years (range, 10~64 years), underwent physical examinations prior to transplantation. The patients had diseases that included AML (n = 109), acute lymphoblastic leukemia (ALL) (n = 69), chronic myelogenous leukemia (CML) (n = 21), MDS (n = 18), aplastic anemia (AA) (n = 15), and other diseases (n = 20).

Sample collection

Ten milliliters of peripheral blood (PB) or bone marrow (BM) was collected from all the donors on the day of transplantation. PB (2 ml) was obtained from all the recipients at 1-month intervals for the first 6 months after allo-HSCT. After this initial period, PB (2 ml) was obtained from the recipients every 3 months, due to the difficulty associated with follow-ups, and missing samples were mainly a result of mortality or disease progression. Sample collection was scheduled at the following times after transplantation: 28 ± 4 days (referred to as the 1 M samples), 58 ± 4 days (2 M), 88 ± 4 days (3 M), 118 ± 4 days (4 M), 148 ± 4 days (5 M), 178 ± 4 days (6 M), 210~270 days (7~9 M), and 300~360 days (10~12 M).

KIR genotyping

Genomic DNA was extracted from PBMCs using a Gentra Puregene Blood Kit (QIAGEN, USA). Genotyping was performed by PCR amplification with sequence-specific oligonucleotide probes (PCR-SSOP), using a KIR Genotyping SSOP Kit (One Lambda, USA). The PCR amplification reactions were performed using a Perkin Elmer GeneAmp 9700 thermal cycler (Perkin Elmer, USA). The data analyses were performed using a Luminex system (USA), and the results were compared with the “Allele Frequency Net Database” (http://www.allelefrequencies.net/kir6001a.asp) (Gonzalez-Galarza et al. 2015). KIR genotyping was performed for all of the enrolled recipient-donor pairs and repeated in the first month after transplantation, and the results revealed that the KIR genotypes of 105 recipients who received specific donor-derived KIR genes transformed into those of their donors during the first month. The distribution of donor-recipient KIR genotypes is listed in Table 1.

Experimental groups

According to the KIR genotype, a recipient with the same KIR genotype as the donor will not display specific donor-derived KIR genes, whereas a recipient who does not share the donor’s KIR genotype will show specific donor-derived KIR genes. Hence, we divided the 252 pairs into two groups: (1) a specific donor-derived KIR gene group (n = 105) and (2) a nonspecific donor-derived KIR gene group (n = 147). The most common combination of specific donor-derived KIR genes and the donor-recipient KIR genotypes can be found in Table 2.

Real-time reverse transcription–qPCR analysis

Total RNA samples from 252 donor-recipient pairs were isolated using the TRIzol Reagent (Invitrogen, USA). One microgram of total RNA was reverse-transcribed using 100 ng of random primers and M-MLV Reverse Transcriptase (Promega, USA). The cDNA templates for PCR were generated by reverse transcription from 2 μg/μl total RNA. Each amplification process involved an individual reaction and was performed in a volume of 25 μl, which included 0.5 μl of the forward primer, 0.5 μl of the reverse primer, 0.3 μl of the TaqMan MGB probe, 12.5 μl of the PCR mix, 7.2 μl of double-distilled water, and 4 μl of cDNA. The 50-ml manually formulated PCR mix contained 10 ml of 10 × buffer, 6 ml of Mg2+ (25 mM), 1 ml of AmpliTaq Gold DNA polymerase (ABI, USA), 0.1 ml of Uracil-DNA glycosylase (Shinegene, China), 1.2 ml of dUTP, 1.2 ml of dATP, 1.2 ml of dCTP, 1.2 ml of dGTP, 1.2 ml of dTTP (Sangon, China), and the rest double-distilled water. For the PCR reactions, the samples were heated at 50 °C for 2 min and 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 60 s. The specificity of the PCR assays was examined through the use of no-template control (NTC) and negative control (NC) samples. No amplification was found in the NTC or NC samples. The PCR assays for all KIR genes and the housekeeping gene ABL were optimized on a LightCycler 480 (Roche, Germany). In each amplification reaction, the correlation coefficient of the standard curve was greater than 0.99, and the slope of the curve was between − 3.3 and − 3.5. The primers and probes were designed with Primer Premier 5.0 (USA) and Primer Express 3.0 (USA), and were synthesized by Thermo (USA). All the sequences of the primers and probes are summarized in Table 3. The copy numbers of the KIR genes are expressed as KIR/ABL ratios and were calculated according to the following formula: (KIR/total ABL ratio × 10,000). This value was defined as the copy number of the KIR gene for every 10,000 ABL gene copies in all of the nucleated cells in the blood.

Statistical analysis

SPSS v20.0 (IBM, USA) was used for the statistical analyses, and graphs were generated using GraphPad Prism 5.0 (GraphPad Software, USA). The KIR gene copy numbers in the transplant recipients are expressed as medians (ranges) and were compared with those in the donors using the Mann-Whitney U test. The contribution of the sex of the recipients and donors was analyzed by a chi-squared test. Fisher’s exact test was utilized to compare the distribution of KIR genotypes among the groups. Continuous variables for AML, ALL, CML, MDS, AA, and other diseases were analyzed by one-way ANOVA.

p values < 0.05 were considered statistically significant, and two-sided tests were used for all the analyses.

Results

Recipient and donor clinical characteristics

Among the 252 enrolled patients, no significant differences were observed among the groups in terms of the age or gender of the recipients and donors. In contrast, significant differences in the total KIR genotype distribution in recipient-donor pairs were found among the groups (p = 0.039). For further cross-group comparisons, we performed a post hoc test with adjusted standardized residuals and considered residuals with an absolute value of more than three to indicate positive significant differences (Agresti 2002). In this manner, we found that the number of BX/AA samples was markedly lower than expected in CML, and higher than expected in the other groups, which included AML, ALL, MDS, and AA. Notably, the distribution of BX/AA donor-recipient pairs did not show significant differences among the AML, ALL, MDS, and AA groups (the absolute value of the residuals was not higher than three). No significant differences in the BX/BX recipient-donor category were observed (p = 0.56).

Analysis of KIR2DL1 and KIR3DL1 expression under the KIR/HLA mismatched model

In the 252 transplantation pairs, 67 pairs exhibited matched KIR/HLA ligands and 185 pairs consisted of mismatched KIR/HLA ligands. According to whether the specific HLA ligands matched between the donors and recipients, we analyzed the expression of KIR2DL1 and KIR3DL1 in the donors and recipients during the first 3 months after transplantation. We found no significant difference in KIR2DL1 expression in the HLA C1/C1, HLA C1/C2 or HLA C2/C2 positive groups (Fig. 1). The same result was found for KIR3DL1 expression among the HLA Bw6/Bw6, HLA Bw4/Bw6 or Bw4/Bw4 positive groups (Fig. 2).

Division of specific donor-derived KIR genes and nonspecific donor-derived KIR genes

Based on the KIR genotyping results, we identified the specific donor-derived KIR genes in the aforementioned experimental group and recorded the high-frequency combinations of specific donor-derived KIR genes in the group (Table 2). The rest of the combinations were not described in detail due to their rarity and small sample sizes. For specific donor-derived KIR genes, the patterns of the donor and recipient KIR genotypes were diverse, and the most frequent combination was donor Bx1 vs recipient AA, followed by donor Bx2 vs recipient AA. In addition, among most of the donor AA vs recipient Bx pairs that were found in this study, specific donor-derived KIRs were not observed except in the case of donor AA vs recipient BB1 or BB6. As for nonspecific donor-derived KIR genes, the highest frequency pair was donor AA vs recipient AA (n = 78, 53.1%), followed by donor Bx1 vs recipient Bx1 (n = 19, 12.9%), and then donor Bx2 vs recipient Bx2 (n = 11, 7.5%).

Expression of KIR genes in the specific and nonspecific donor-derived groups

The median expression of KIR2DS1 in the specific donor-derived group exceeded in the donors in the first month after allo-HSCT and reached its peak value of 2372.2 (range, 817.5~6828.9) in the second month posttransplantation (Fig. 3a). However, the expression decreased after this peak until it approached the donor level at 1-year posttransplantation. Interestingly, the median expression of KIR2DS1 in the nonspecific donor-derived group was similar to that in the specific donor-derived group and reached its peak value of 2722.2 (range, 938.9~7865.9) in the second month posttransplantation. No significant differences in expression were found between these two groups at each time point.

The median expression of KIR2DS3 in the specific donor-derived group in the first month posttransplantation was 2082.3 (range, 198.9~4727.9) (Fig. 3b), and the expression gradually decreased until reaching a level that was approximately equal to the donor level within 1-year posttransplantation. In the nonspecific donor-derived group, the level observed in the first month posttransplantation was 1955.7 (range, 854.2~3434.3), and the observed trend was similar to that found in the specific donor-derived group. No significant differences were found between the two groups at any time point. The expression of KIR2DS3 in the nonspecific donor-derived group was not significantly different from that in the donors during the study period, whereas the expression in the specific donor-derived group in the second month posttransplantation was higher than that observed in the donors (1486.8 (range, 590.1~4351.0) vs 910.7 (range, 282.4~2580.8), p = 0.04)).

The median expression of KIR2DS4 in the two groups gradually increased until the third month posttransplantation. In the second month, the expression of KIR2DS4 in the specific donor-derived group was lower than that in the nonspecific donor-derived group (1736.0 (range, 1120.4~4085.2) vs 3027.78 (range, 913.3~7610.1), p = 0.12)) (Fig. 3c). The highest KIR2DS4 expression in the specific donor-derived group was 3409.1 (range, 1670.9~4090.9), which was lower than that in the nonspecific donor-derived group (3767.2 (range, 1198.3~8985.8)), but this difference was not significant.

The median expression of KIR3DS1 in the specific donor-derived group continuously increased during the study period, and in the third month, the median KIR3DS1 expression in the specific donor-derived group reached a value of 3809.2 (range, 313.4~7158.5) (Fig. 3d), which was higher than the expression level detected in the donors and recipients at all tested timepoints. During the rest of the study period, decreases in the KIR3DS1 expression levels with small intermittent fluctuations were observed in both the specific donor-derived and nonspecific donor-derived groups. No significant differences were observed between the two groups.

Comparison of iKIR and aKIR expression

The expression of iKIRs, including KIR2DL1, KIR2DL2, and KIR3DL1, continuously increased until the third month after transplantation, and significant differences in their expression, particularly that of KIR2DL1 and KIR3DL1 in the second and third months posttransplantation (p < 0.0001), were found compared with the donor level (Fig. 4a). In the third month, the highest expression was found for KIR3DL1 (2361.0 (range, 386.1~6729.8)), followed by KIR2DL2 and KIR2DL1, the latter of which exhibited a value of 947.7 (range, 205.2~3743.6). The expression of KIR2DL2 in the third month was clearly higher in the recipients than in the donors (1762.4 (range, 354.8~5919.3) vs 714.3 (range, 355.9~2991.3, p = 0.01)). In contrast, in the first month posttransplantation, the KIR gene levels in the recipients were not significantly different from those in the donors, although the median expression levels were higher than those in the donors. Starting in the third month, the frequencies of KIR2DL1, KIR3DL1, and KIR2DL2 decreased until they reached the donor levels 1 year after allo-HSCT. The same trends were observed for the reconstituted copy numbers.

The median expression levels of KIR2DS3 and KIR2DS5 reached their peak values in the first month posttransplantation, whereas KIR2DS1 expression reached its peak in the second month and the highest KIR2DS2, KIR2DS4, and KIR3DS1 expression levels were detected in the third month. The expression of the target KIR genes gradually increased to the peak value and then began to decrease, until reaching the donor level within 1 year after transplantation, and the same trend was found for all the reconstituted KIR genes that were studied. The copy numbers of most aKIRs in the recipients were not significantly different from those in the donors; however, the median expression of KIR2DS3 in the recipients in the second month was higher than that in the donors (1486.8 (range, 590.1~4351.0) vs 910.7 (range, 282.4~2580.8), p = 0.04)) (Fig. 4b), and the median expression of KIR2DS2 in the recipients in the second month was higher than that in the donors (1169.7 (range, 435~3596.7) vs 640.7 (range, 430.2~1723.4), p = 0.0043)) (Fig. 4b).

The copy numbers of iKIRs were different from those of specific donor-derived aKIRs. The expression of iKIRs remained markedly elevated in the recipients between the first and third months after transplantation, whereas the expression of aKIRs in the recipients was higher than that in the donors by only a nonsignificant margin. The peak values for KIR2DL1, KIR3DL1, and KIR2DL2 were detected in the third month posttransplantation, whereas the peak KIR2DS3 and KIR2DS5 copy numbers were obtained in the first month, and the peak value for KIR2DS1 expression was reached in the second month. However, the expression of both iKIRs and aKIRs reached the donor levels within the first month after transplantation gradually increased to reach their peak values and then decreased to return the donor levels within 1 year.

Discussion

NK cells integrate signals triggered by interaction of target cell ligands with an array of activating and inhibitory KIRs after allo-HSCT. Specific donor-derived KIR genes can be distinguished from nonspecific donor-derived ones through KIR genotyping of donor-recipient pairs. In this study, we observed that 105 of the 252 enrolled donor-recipient pairs possessed specific donor-derived KIR genes. It was a shortcoming that no detection method for the expression of specific donor-derived KIR genes was performed in the previous studies. One of the most common methods used to measure the expression of KIR genes is flow cytometric analysis, which can hardly distinguish the expression of aKIRs and iKIRs, let alone that of specific donor-derived KIR genes. We previously dynamically detected the expression of KIR2DL1/3DL1 protein and mRNA in NK cells (Hu et al. 2017). The median protein expression level of KIR3DL1 in the posttransplant recipients in the first month exceeded the donor level and reached the highest value in the third month, and this finding potentially implicates the reconstitution of NK cells in the expression of membrane protein. The comparison of the recovery trend of KIR3DL1 protein and mRNA revealed that the mRNA expression of KIR3DL1 was in line with that of the membrane protein, even though the recovery of KIR2DL1 mRNA expression occurred earlier than that of the protein, and this delay might be due to the interaction between KIR2DL1 and HLA class I ligands and the polymorphisms in different KIR repertoires. Giebel et al. (2010) evaluated the recovery of the expression of NK cell iKIRs through cytometric analysis. Gallez-Hawkins et al. (2011) studied the mRNA expression of two aKIRs, KIR2DS4 and KIR2DS2, during the postinfection phase, and various novel KIR genotype methods have been used to study polymorphisms of KIR genes (Chaisri et al. 2018; Kitpoka et al. 2016; Wang et al. 2012; Yao et al. 2019). Chen et al. (2009) designed a quantitative KIR RNA type assay and found leukemia patients exhibited lower KIR expression before transplantation than healthy donors. These researchers also hypothesized that mobilization might transiently enhance donor KIR expression but does not affect the balance between iKIRs and aKIRs. Denis et al. (2005) found that the delayed appearance of KIR transcripts posttransplantation might be associated with lower absolute numbers of CD45+, CD3−, CD16+, CD56+ NK cells and a higher risk of acute GVHD. These results suggest that KIR expression undergoes kinetic changes that might influence the process and outcome of transplantation. However, our predictions call for an innovative method with better quantification and dynamics than other previously developed tests. Hence, we designed a method for measuring the mRNA expression of KIR genes based on genetic cloning techniques and applied for the domestic patents on our method. The RT-qPCR assays for all the KIR genes and the ABL housekeeping genes were optimized simultaneously because ABL is expressed at similar levels in healthy individuals and patients (Beillard et al. 2003). Considering the potential effect of KIR/HLA mismatch on mRNA expression, we analyzed the mRNA expression of KIR2DL1 and KIR3DL1, and found no significant difference due to KIR/HLA mismatch. Hence, the mRNA expression patterns obtained in our study could be used for distinguishing specific and nonspecific donor-derived KIR genes. For instance, a donor Bx1 vs a recipient AA, specific donor-derived KIR genes allowed for detecting mRNA expression can be regard as KIR3DS1, KIR2DL5, KIR2DS5, and KIR2DS1. According to our research, the expression of each specific donor-derived KIR gene was in line with that of the nonspecific donor-derived KIR gene. Therefore, we also evaluated the recovery pattern of the expression of aKIRs and iKIRs at different timepoints after transplantation.

The specific donor-derived KIR genes were mostly aKIRs, and only a small proportion were iKIRs. KIR3DL1 and KIR2DL1 account for the majority of specific donor-derived iKIRs. The ligands of these two proteins have been thoroughly studied and discussed, yet the research on that of aKIRs was still unknown (Anfossi et al. 2006; Boudreau et al. 2016; Fernandez et al. 2005; Heidenreich and Kroger 2017; Kim et al. 2005; Yawata et al. 2008). The main specific donor-derived aKIRs are KIR2DS1, KIR2DS3, KIR2DS5, and KIR3DS1. Based on the findings obtained in this study, we established the chronology of the peak values of aKIRs after transplantation. Specifically, we found that the expression of KIR2DS3 and KIR2DS5 peaked in the first month after transplantation, and KIR2DS1 expression reached its highest level in the second month, and that the highest values for KIR2DS4 and KIR3DS1 expression were detected in the third month. Moreover, each aKIR showed a different expression level at every time point after transplantation, and their expression was dynamic but relatively stable. Therefore, the aKIRs were derived from the donors and could be used for comparison with pretransplantation results and for clinical follow-up. The correlation between mRNA expression changes, and manifestations are being analyzed in our ongoing work. In addition, the expression of specific donor-derived KIR2DS3 and KIR2DS2 in the second month posttransplantation was significantly different from the donor levels. These results were likely obtained due to the small sample size and the inequitable genotype distribution of KIR2DS2 in the Chinese Han population. More specific samples are necessary to characterize the expression level in detail. Compared with the aKIR repertoires, iKIRs involve not only definite ligands but also different arrangements in the genomic structure of KIRs (Barten et al. 2001; Martin et al. 2000, 2004). We found that the iKIR repertoires exhibited a different recovery trend from the aKIRs, and their trend involved a more rapid development and higher expression level posttransplantation relative to the donor level. Interestingly, all the examined iKIRs reached their highest expression levels in the third month posttransplantation.

Rapid recovery indicates a significant role for NK cells in the early phase after transplantation. It is well known that allo-HSCT faces two difficulties that require breakthroughs, i.e., the restoration of hematopoiesis and the reconstitution of the immune system. In this study, we focused on immune reconstitution and designed a method for quantitatively measuring the expression of specific donor-derived KIR genes. The dynamic changes in the expression of these KIR genes could serve as a biomarker for predicting and evaluating NK cell functional recovery after transplantation. As previous study indicated that a beneficial alloreactive NK cell response can be induced in recipients who lack a KIR ligand for donor-derived educated KIRs (Cooley et al. 2018). Several studies have hypothesized that a missing HLA ligand for donor inhibitory KIRs is associated with no-longer-suppressed inhibitory signals, enhanced cytotoxicity, and increased affinity between activating receptors and their ligands. We are studying, whether under discrepant KIR/HLA models, the levels of KIR mRNA expression exert an influence on the occurrence of GVHD, overall survival, and relapse-free survival in patients with malignant hematologic diseases.