杀伤细胞免疫球蛋白样受体基因及其配体相合或错配对单倍相合骨髓移植效果的影响

作者：段连宁， 韩红星， 刘静， 阎洪敏， 朱玲， 薛梅， 丁丽，朱培瑜， 王恒湘， 纪树荃    作者单位：空军总医院血液科，北京 100036

【摘要】  本研究分析杀伤细胞免疫球蛋白样受体（killer cell Iglike receptor, KIR）及其配体分子相合与否对单倍体相合骨髓移植效果的影响。 分析了74例KIR基因及其配体分子HLACw的分布频率和特点，同时比较KIR分子配体缺失与否对单倍相合骨髓移植患者总体生存、无病生存、GVHD发生以及复发等的影响。结果表明： 19个KIR基因表型中2DL1、2DL4和3DL23在所有个体100％分布，其他高频率分布的还有3DP1（98.6％）、2DP1（98.6％）、3DL1（97.3%）、2DL3（97.3%）；抑制型基因分布频次是活化型的1.37倍；所有单倍体都含有2DL1、KIR3DL2、3DL3和2DL4，其中每个单倍体中均含有2DL2和/或2DL3。 HLAC 14个等位基因中Cw7分布频率最高为37.8%； KIR2DL2/2DL3识别配体Group2(HLACw1,3,7,8,13,14)组占43.2％。32例单倍相合骨髓移植中KIR基因错配发生比例为43.8%，其中9/14例为2DL不相合、5/14为2DL2或3DL1不相合；46对单倍相合骨髓移植中HLACw全相合者29例，不相合者14例，HLACw发生完全错配比例为30.4%，全相合比例为63.4%。14例KIR基因不相合和13例KIR基因相合移植病例中，KIR基因相合组生存率高于KIR基因不相合者(p=0.032)； 17例KIR分子配体HLACw不相合移植患者的DFS明显高于24例相合者（p=0.024）。按供受者KIR配体是缺失的GVHD矢量组生存率高于非GVHD矢量组（p=0.015）。急性重症GVHD发生与活化型KIR2DS1/2DS2有关，2DS1/2DS2不相合组高发急性重症GVHD同时复发减低，而2DS1/2DS2相合组低GVHD但是复发增多。 14例KIR配体不相合髓系白血病患者骨髓移植后1例复发死亡，而12例KIR配体相合组患者有4例复发死亡。结论： 单倍体相合骨髓移植的主要特点就是HLA不全相合，KIR基因表型不一致性及其配体缺失也是其主要免疫学特征，供者型KIR配体的缺失与移植效果有密切关系，在单倍相合移植中分析KIR基因及其配体对于供者选择和判断预后有重要意义。

【关键词】  杀伤细胞免疫球蛋白样受体

    Impact of Incompatible Killer Cell Immunoglobulinlike  Receptor and Its Ligand on the OUtcome of Haploidentical     Compared with HLA identical HSCT or BMT, haploidentical BMT faced much stronger reactions between donor and recipient especially of the graft versus host reaction (GVHR). The current opinion believed that the occurrence of GVHD was helpful in preventing the relapse of malignant hematological disease because the antileukemia effect (GVL) induced by donor derived T cells was dependent on the coincidence of GVHR when failing to clarify the difference between GVHR and GVL. The most popular measure to promote engraftment and inhibit GVHD is still by the blockade of T cell activation or by T cell inhibition/depletion in vivo. But the GVL effect induced by donor derived T cells is also weakened through above measures. A novel research about the NK immunologicallike receptor found that NK cells can exert the homogeneous antileukemia effect through an overall different mechanism. Donorderived homoreactive NK cells played a key role in improving the effect of BMT［1］.    More research followed showed that the first recovered NK cells after BMT had some thing to do in preventing GVHD, initiating and promoting the engraftment and inducing the antileukemia effect. The KIR expressed on the surface of NK cells played an important role during the process［2-5］.  KIR is a member of immunologicallike superfamily expressed on NK cells which could be classified into two groups namely inhibitory group (possessing a cytoplasmic tail bearing immunoreceptor tyrosinebased inhibitory motifs, ITIMs) and activating group (through adaptor protein to generate intracellular signals via immunoreceptor tyrosinebased activating motifs, ITAMs) which can recognize the correspondent HLAI ligand to exert immunological effect. If the KIR ligand was missing between donor and recipient   （missing self）   the NK cells could display heterogeneity even homogeneity reactive potential to facilitate GVL and decrease relapse, GVHD and transplantation failure and prolong overall survival (OS) rate［1，2］. Therefore the genotyping of KIR and its ligand had been given more and more emphasis recently by the scientists engaging in BMT.

    Up to now 19 kinds of KIR genotypes were  identified. Among them the inhibitory KIR genotypes were paid much more attention because of their relations with GVL when incompatible with their ligand although activating KIR worth research in the way connected with GVHD. The role of KIR were valuable only in combination with their HLAI ligands, that  was helpful to estimate the BMT effects by analyzing the compatibility of KIR and their ligands.  Most  related research focused on the HLA matched BMT to evaluate the incompatibility of KIR ligand, but few was done about haploidentical BMT［6-8］. This study  analyzed the genotype of KIR and its distribution frequency as well as their impact to BMT effect whether  compatible or not between donor and recipient, that  provides a guide for the prognostic and treatment procedure.

    Material and Methods

    Patients and Donors

    74 patients underwent haploidentical BMT in the General  Hospital Air Force of   PLA  during December 2002 to  December 2005,  so that an  intact data about KIR genotype and its HLACw ligand was analyzed.  Among them, 32 pairs were suitable for further compatibility research between KIR and its ligand. Table 1 showed the general conditions of the patient whose donor were all health relatives with mismatched HLA.

    Genotyping of HLACw and KIR

    HLACw type was determined by sequencespecific polymerase chain reaction (PCRSSP) kit provided by PELFREEZ company and all the 14 HLACw geno

    Table 1. Characteristics  of 32 patients undezgimg heploidentical BMT

    Mean age, yrs（range）25.3±11.9(3－46)  Sex (male/female)19/13Disease    AML8  CML15  ALL4  Lymphoma2  MDS1  Others2Conditioning    CTX+AraC+TBI+ATG+antiCD2531  FLU+Bu+CTX+ATG+antiCD251GVHD prophylaxis  MTX+CSA+MMF32BMT direction   parent→children14  sibling→sibling15  children→parent3Status of haploidentity for HLAA,B,DR    3/6 mismatch20  2/6 mismatch12Missing or not of KIRligand for recipient   KIRligand（HLACw）presence14  KIRligand（HLACw）absence18Note: AML— acute myelogenous leukemia, CML— chronic myelogenous leukemia, ALL— acute lymphocyte leukemia, MDS—myelodisplastic syndrome, CTX— cyclophosphamide, AraC—cytarabine, TBI— total body irradiation, FLU—fludarabine, Bu—carmustine, MTX—methotrexate, CsA—cyclosporin A, MMF—mycophenolate mofetil.

    types could be specified.  KIR genotype was also determined  by PCRSSP kit provided by the same company which can distinguish 19 genotypes including 2DL1, 2DL2, 2DL3, 2DL4, 2DL5A, 2DL5B,2DS1,2DS2,2DS3, 2DS4,2DS4,2DS5,3DL1,3DL2,3DL3,3DS1,2DP1,3DP1 and 3DP1.

    The grouping of KIR and Its Ligand

    All of the 19 genotypes of KIR and 14 genotypes of HLACw were found in the 74 cases whereas only the distribution characteristics, frequencies and haplotype inheritance of KIR and its ligand were analyzed. Of the 32 pairs underwent haploidentical BMT two groups were established based on compatibility for KIR genotype and matching state for its ligand respectively. Then the absence or presence of KIR ligand corresponding to the donor was regarded as the missing ligand group (GVHD group) and nonmissing group (non GVHD). At last we analyzed the role of activating KIR whether or not compatible. The above discrepancies in terms of the overall survival(OS), diseasefree survival( DFS), relapse and acute heavy GVHD  were evaluated.

    BMT and Conditioning Regimens

    BMT, conditioning regimens and GVHD prophylaxis were based on the references［9］ with an exception for a thalassemia patient conditioned by FLU+Bu+CTX+ATG+anti CD25.

    Statistical Analysis

    The average and standard test was used for frequency material while the descriptive method was used for the counting of GVHD and relapse cases due to the shortage of total number. The KaplanMeier method was used in analyzing OS and DFS. All the data were calculated by SPSS 10.0 software.

    Results

    Distribution frequencies of KIR genotypes

    Among the 19 KIR genotypes currently nominated  KIR2DL1, KIR2DL4 and  KIR3DL23 could be  detected in all the cases. Other high frequency genotypes included KIR3DP1（98.6％）, KIR2DP1（98.6％）,KIR3DL1（97.3％）and KIR2DL3（97.3％）. The low frequency genotypes were 2DS1（33.8%）,3DS1(32.4%),2DL5(31.1%),2DS4(25.7%) and 2DS5(23.0%) while all the others were lower than 15% (Figure 1). The frequency for inhibitory type (L) was 475 times and activating type (S) 348 times. The former was 1.37 times of the latter.

    Figure 1. Frequencies of KIR genotype for the 74 related cases.

    The comparison of KIR and its ligand between donor and recipient for the 32 pairs showed that KIR2DL1, KIR3DL2, KIR3DL3 and KIR2DL4 were found in all haplotypes and at least one genotype of KIR2DL2 and/or KIR2DL3 exists in all haplotypes with a high frequency for KIR2DL3. All the genotypes mentioned above belonged to inhibitory type. The distribution of activating types were not balanced.

    Distribution Characteristics of KIR2DL1/2DL2/2DL3 Ligand  HLACw

    Among the 14 genotypes found in the test the HLACw7 was the most popular (37.8%) and other high frequent genotypes were Cw3（29.7%），Cw1（28.4%），Cw4(25.7%)，Cw6(25.6％), Cw8(21.6). Other genotypes with low frequencies were distributed by Cw12(12.2%), Cw15(8.1%), Cw2(5.4%), Cw5(2.7%) and Cw14（1.4%）. HLACw16,17 and 18 were not detected for this group. Among the positive genotypes 32 cases belonged to the group 2 (HLACw1,3,7,8,13,14) distinguished by KIR2DL2/2DL3; 13 cases belonged to the group 1 (HLACw2,4,5,6,12,15,17) distinguished by KIR2DL1; 29 cases contain ligands distinguished by the both groups. The relations between KIR2DL2/2DL3 and its ligand group 2 should be paid more attention because of its high frequencies.

    Analysis of inhibitory KIR genotypes and its mismatched ligands between donor and recipient in haploidentical BMT

    18 cases were compatible for the inhibitory KIR genotype among the 32 cases of haploidentical BMT and the others were not. The incompatibility ratio was 43.8% while 9/14 by KIR2DL site and 5/14 by KIR2DL2 OR KIR3DL1.

    29 cases were matched for HLACw among the 46 pairs of donor and recipient with complete typing data for analysis while 14 were mismatched. The 3 left were permitted mismatch by HLACw group 1 and group 2. The mismatching ratio was 30.4% and matching ratio was 63.4% while permitted mismatch was 6.5%.

    Impact of compatibility of KIR genotype to the survival after haploidentical BMT

    The comparison between the 14 compatible cases and 13 incompatible cases after haploidentical BMT showed that the former had a higher survival rate than the latter (p=0.032) which was showed in Figure 2.

    Impact of KIR ligand to the survival after haploidentical BMT

    The comparison between 24 cases of matched KIR ligand and 17 cases of mismatched KIR ligand in terms of survival was showed in Figure 3.

    Figure 2. Relationship between KIR genotype and survival after haploidentical BMT.

    Figure 3.  Relationship  between KIRligand HLACw and survival after haploidentical BMT.

    Impact of GVH direction reactions induced by KIR to the survival after haploidentical BMT

    Two groups were set according to the presence or absence of KIR ligand. The first including 10 cases was ligand absence group by GVH direction ( the ligand HLACw recognized by donor KIR was missing on the recipient) and the second including 11 cases was ligand absence group by HVG direction ( the ligand HLACw recognized by recipient KIR was missing). Actually it was grouped by the homogeneity reactive potentials between donor and recipient NK cells.  The group of GVH direction could initiate homogeneity reaction while the other group could not. The results showed that the survival rate was higher for the former group which was showed in Figure 4.

    Impact of activating KIR compatible or not with GVHD and relapse after haploidentical BMT

    Activating KIRs  mainly consisted of KIR2DS15 and 3DS1, among which 2DS1 and 2DS2 were important. The GVHD analysis done by comparing the 13 compatible cases with the 14 incompatible cases after haploidentical BMT showed that 2DS1 and 2DS2 had something to do with the acute severe GVHD. The acute severe GVHD was higher and relapse was lower for the incompatible group while it was opposite for the compatible group(Figure 5). However the result was drawn from limited cases and should be verified by further accumulated data. Of 14 cases suffering from myeloid leukemia with mismatched KIR ligand one patient died from relapse after haploidentical BMT while 4 patients died from relapse for the 14 cases of matched group. About lymphoid leukemia none relapsed for the mismatched group while both of 2 cases for the matched group died from relapse.

    Figure 4.  Impact of GVH direction homogeneity reaction potentials induced by KIR to the survival after haploidentical BMT.

    Figure 5.  Relapse for patients with acute myeloid leukemia after BMT of matched and mismatched KIR ligand.

    Discussion

    KIR was firstly found on NK cells that could combined specifically with target ligand HLAI such as KIR2DL1 with HLACw2、4、5、6、12、15( group 1) and KIR2DL2/2DL3 with HLACw1、3、7、8、13、14（group 2）. When the ligand recognized by KIR was absent the target cell would be killed ("missing self" mechanism) ［10］.  The presence of KIR also played an important role in keeping homeostasis by preventing self attack from activated immunological cells.  The effect of GVHD and GVL induced by NK cells was dependent on the KIR and its ligand.  In the research for partially matched BMT about CML patients Albi found［11］ CTL from the donor highly expressing inhibitory KIR could restrain GVHD while remaining GVL. The low aGVHD was related to highly expressed KIR after HSCT. The expression of KIR in T cells in the early stage after BMT might belong to a molecular reaction to give an immunological protection or weaken the immunological assault.

    The distribution character of KIR showed that KIR3DL2、3DL3、2DL4 and 2DL were inherited by linkage. Many KIRs could be expressed simultaneously by one person and both activating and inhibitory KIRs could exist in the same repertoire. What worth mention was KIR2DL4 that had both ITAM and ITIM functions.  The higher expression of inhibitory KIR than activating KIR was important in keeping immunological homeostasis. Therefore the role of inhibitory KIR should be pay more attention in the alloreactions after homogeneous transplantation.   KIR worked by combining with a group of ligands but not a special one. HLACw could be divide into two groups (group 1 and group 2) according to their recognition by KIR.  When both inhibitory and activating KIR could recognize the same antigen clusters the inhibitory effect would predominant because the former KIR had higher affinity with the ligand which was also the biochemical mechanism to prevent self assaulting by NK cells. It seemed that the group 2（HLACw1、3、7、8、13、14）was more common than group 1 during our populations. 63.4% of patients was KIR compatible and 30.4% was incompatible that underwent haploidentical BMT. Therefore it was common for haploidentical BMT with a matched HLACw which could serve as a reference in the selection of donor of mismatched KIR ligand.

    The results showed that the meaning of incompatible      KIR genotypes was completely different with that of mismatched HLACw.             The mismatched group of KIR ligand had a higher overall survival rate. However the incompatible group of KIR genotype had a lower overall survival rate which seemed contrary to the above results. The reason for this was that KIR genes and HLACw genes were located on different chromosomes and inherited independently by Mondel rules. The epitopes of KIR were independent of the expression of KIR ligand.  The incompatibility of KIR genotypes between donor and recipient did not necessarily means the mismatch of its ligand because  there were no relations in terms of their genotypes. Even HLA completely matched individuals might had different KIR genotypes［12］.  Just based on above theory, we could not only focus on the KIR genotypes in analyzing the relations between KIR and BMT and should also consider the role of its ligand. KIR could combined with many HLA class I molecules with HLACw as the most advantaged one［13］. Mismatched HLACw could served as the marker for a long survival. The likely mechanism was that the NK cells could induce antileukemia alloreactions when KIR was compatible to prevent relapse and at the same time the alloreactions could kill host APC cells to minimize the activation of T cells which could decrease acute GVHD. The data in this study also proved that the DFS was higher in the GVH group (missing ligand of the recipient for the donor KIR) than in the nonGVH group and verified the theory that the potential alloreactivation induced by donor KIR could letdown acute severe GVHD and  improve survival［14］. What important was this kind of alloreactivation of NK cells had no response to normal host tissues because of the lack of the activating KIRs which was also one of the reason of not initiating GVHD by NK cells. However there were opposite reports［15］. A research group from Israel had transfused pure alloreactive NK cells to treat the relapse after BMT and had approved that NK cells could prevent relapse without GVHD［16］.

    It was more valuable to explore the relations between activating KIR and GVHD compared with inhibitory KIR. A multicenter research to 174 cases of HLA matched HSCT from France showed that the activating KIR was related to a high rate of GVHD and the incompatibility of activating KIR should be avoid in HSCT. They also advised to do KIR typing in selecting the proper donor for HSCT［17］. In this study only one out of 13 cases with compatible KIR2DS1/2DS2 developed acute severe GVHD while it was 5 out of 13 for incompatible group. It was interested to find that the incompatible group of KIR2DS1/2DS2 developed high GVHD and low relapse while the compatible group developed low GVHD and high relapse. We could not drew the conclusion whether the missing of activating KIR such as  2DS1、2DS2、2DS3、2DS4 and 2DS5 would impact the OS and DFS currently because of the shortage of evidence based on the clinical data home and abroad［6］. The reports on the relations of incompatible KIR with high GVHD were not consistent. The main stream attitude believed that NK cells had no relations with high acute severe GVHD otherwise could contri bute to a high survival when incompatible or being mismatched ligand.  It was worth further study for the immunological function of activating KIR in haploidentical BMT from the angle of immunological mechanism of KIR recognition.

    For all the patients suffering from myeloid leukemia only one out of 14 cases with mismatched KIR ligand died from relapse while 4 out of 14 with matched KIR ligand died from relapse after haploidentical BMT. It supposed to be more efficient to exert alloreactive GVL effect of the NK cells than T cells  because of their different recognition and killing mechanism.  As we all knew, it was still difficult to control the relapse for the patients suffering from lymphocytic leukemia after BMT because of limited measures. However in this study none of the 4 patients with mismatched KIR ligand relapsed comparing with both of the 2 matched patients died from relapse. Some reports proved that the AML patients( including MDS) were better than severe CML and ALL in terms of OS, DFS and lower rate of relapse when missing the KIR ligand for donor［6］. Although the data was not substantial, it was still a novel to prevent the relapse of leukemia by selecting the donor source of immunological cells which had mismatched or missing KIR ligand when considering the incapability of traditional BMT in controlling the relapse of leukemia only by T cells. Actually Slavin［16］ had achieved inspiring results by transfusion of pure NK cells activated by IL2 to relapsetending or relapsed patients. In fact the antirelapse function by donor peripheral MNC transfusion was dependent on the mismatch of KIR ligand and which had highlighted a future for relapsed patients. By the way the influencing factors of OS, DFS, GVHD and relapse after haploidentical BMT were very complicated especially when the HLAA, HLAB and HLADR genotypes were 3/6 or 2/6 mismatched. However they did not bother us in analyzing the role of KIR and its ligand HLACw as an independent factor  especially after the inhibition of T cells function by the adding of antiCD25 in this study.

    In summary, what haploidentidcal BMT means was partially matching. Then it was more common for  incompatible KIR and mismatched or missing ligand than matched BMT.  The predominance of KIR and its ligand should be attached more importance in the role of GVHD controlling and GVL enhancing. KIR ligand analysis might as well be served as one of the main concern in the selection of proper donor before BMT and the immunotherapy treatment after BMT［7，17］.

    Acknowledgment: We should express our thanks to  Director CAI Qing of Centeral Laboratory in General Hospital of Air Fore of PLA in his great help in the typing of HLA and KIR.

【参考文献】1Ruggeri L, Capanni M, Casucci M, et al. Role of natural killer cell alloreactivity in HLAmismatched hematopoietic stem cell transplantation. Blood, 1999; 94:333-339

2Ruggeri L, Capanni M, Urbani E,et al. Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants. Science, 2002;295(5562):2097-2100

3Tanaka J, Mori A, Ohta S, et al. Expression of HLACspecific natural killer cell receptors (CD158a and CD158b) on peripheral blood mononuclear cells after allogeneic bone marrow transplantation. Br J Haematol, 2000;108:778-783

4Small TN, Avigan D, Dupont B, et al. Immune reconstitution following Tcell depleted bone marrow transplantation: effect of age and posttransplant graft rejection prophylaxis. Biol Blood Marrow Transplant, 1997; 3:65-75

5段连宁，陈纯，黄绍良等. 免疫活性细胞表达杀伤细胞抑制性受体与造血干细胞移植后移植物抗宿主病的关系. 中国实验血液学杂志，2003； 11：625-632

6Hsu KC, KeeverTaylor CA, Wilton A,et al. Improved outcome in HLAidentical sibling hematopoietic stemcell transplantation for acute myelogenous leukemia predicted by KIR and HLA genotypes. Blood, 2005 ;105:4878-4884

7Cook MA, Milligan DW, Fegan CD, et al. The impact of donor KIR and patient HLAC genotypes on outcome following HLAidentical sibling hematopoietic stem cell transplantation for myeloid leukemia. Blood, 2004;103:1521-1526

8Leung W, Iyengar R, Turner V, et al. Determinants of antileukemia effects of allogeneic NK cells. J Immunol, 2004;172:644-650

9Ji SQ, Chen HR, Wang HX, et al. GCSFprimed haploidentical marrow transplantation without ex vivo T cell depletion: an excellent alternative for highrisk leukemia. Bone Marrow Transplant, 2002; 30:861-866

10Karre K, Ljunggren HG, Piontek G, et al. Selective rejection of H2 deficient lymphoma variants suggests alternative immune defence strategy. Nature, 1986; 319 (6055): 675-678

11Albi N, Ruggeri L, Aversa F, et al. Natural killer (NK)cell function and antileukemic activity of a large population of CD3+/CD8+ T cells expressing NK receptors for major histocompatibility complex class I after "threeloci" HLAincompatible bone marrow transplantation. Blood, 1996; 87:3993-4000

12Dupont B, Hsu KC. Inhibitory killer Iglike receptor genes and human leukocyte antigen class I ligands in haematopoietic stem cell transplantation. Curr Opin Immunol, 2004;16:634-643

13Farag SS, Fehniger TA, Becknell B, et al. New directions in natural killer cellbased immunotherapy of human cancer. Expert Opin Biol Ther, 2003;3:237-250

14Giebel S, Locatelli F, Lamparelli T, et al. Survival advantage with KIR ligand incompatibility in hematopoietic stem cell transplantation from unrelated donors. Blood, 2003; 102:814-819

15Bishara A, DeSantis D, Witt CC, et al. The beneficial role of inhibitory KIR genes of HLA class I NK epitopes in haploidentically mismatched stem cell allografts may be masked by residual donoralloreactive T cells causing GVHD. Tissue Antigens, 2004; 63:204-211

16Slavin S,Shapira MY,Morecki S, et al. Immunotherapy for resistant hematologic malignancies using matched or mismatched rIL2 activated donor lymphocytes positively selected for CD56+ after allogeneic stem cell transplantation for allogeneic cell therapy without GVHD. Blood, 2003;102: Abstract 5329

17Denis L, Sivula J, Gourraud PA, et al. Genetic diversity of KIR natural killer cell markers in populations from France, Guadeloupe, Finland, Senegal and Reunion. Tissue Antigens, 2005; 66:267-276