Extended loci histocompatibility matching in HSCT—Going beyond classical HLA

3.1 MICA 3.1.1 Structure

The major histocompatibility complex (MHC) class I chain-related A (MICA) is a single chained glycoprotein that is located approximately 46 kb centromeric to HLA-B (Bahram et al., 1994; Leelayuwat et al., 1994). This proximity results in a strong linkage disequilibrium between MICA and HLA-B (Collins, 2004; Stephens, 2001). The amino acid sequence of MICA shows up to 36% homology to that of HLA class I, subsequently resulting in a molecular structure that resembles that of HLA class I proteins. However, MICA does not associate to the β2-microglobulin nor does it present peptides (Bahram et al., 1994; Leelayuwat et al., 1994). MICA is a highly polymorphic gene, with the highest degree of polymorphism being clustered in exons 2–5. Currently, 224 MICA alleles and 104 encoded proteins have been described (www.ebi.ac.uk/ipd/imgt/hla/stats.html, release 3.42, 2020-10). In addition to a variety of exon located SNPs, a variation of short tandem repeats (STRs) and an insertion of a single guanine-insertion following the fifth STR within exon 5 also account for the high polymorphism observed between MICA alleles (Mizuki et al., 1997). It is of note that more than 50% of the Caucasian population bears this G-insertion (MICA-A5.1), which results in a premature stop codon and consequently in a truncated MICA protein (Petersdorf et al., 1999; Zhang et al., 2001). Such truncated MICA proteins are being atypically expressed on the cell surface and potentially increase the levels of soluble isoforms of the MICA protein (sMICA) into the extracellular space (Mizuki et al., 1997; Tamaki et al., 2009).

3.1.2 Expression

The expression of MICA molecules is characterized by a limited tissue distribution. The MICA proteins are constitutionally expressed on the surface of the gastrointestinal epithelium (Stephens, 2001). Cell surface expression has also been demonstrated for endothelial cells, fibroblasts and monocytes (Groh et al., 1996; Zwirner et al., 1999), as well as, within a specific stimulating environment, on activated T cells (Molinero et al., 2002, 2003, 2004). MICA is considered as a cell stress marker, as its expression is inducible by heat, viral infection, inflammation and DNA damage (Gasser et al., 2005; Groh et al., 1996, 2001; Schreiner et al., 2006). As MICA is also expressed on different solid tumours (Groh et al., 1999), melanomas and leukemic T-cell lines (Pende et al., 2001), it has been suggested that MICA expression can be also induced upon neoplastic transformation, mediated by transcription factors (Andresen et al., 2007) and nuclear receptors (Jinushi et al., 2003).

3.1.3 MICA antibodies in HSCT

It has already been shown that MICA antigens are able to induce an antibody-mediated alloreaction that may kill target cells in the presence of complement (Zou et al., 2002). That said, in the context of solid organ transplantation, MICA antibodies are possibly associated with graft rejection (Zou et al., 2006; Zou & Stastny, 2009). As for HSCT, it has been reported that preformed MICA antibodies prior to transplantation may correlate with a decreased probability of developing chronic GvHD, whereas post-HSCT formed MICA antibodies may increase the risk of relapse (Boukouaci et al., 2009).

3.1.4 Soluble MICA in HSCT

MICA acts as a ligand to the NKG2D receptor on NK cells, γδ T cells and αβ CD8+ T cells (Bauer et al., 1999). NKG2D is an activating receptor that triggers cytotoxicity (Billadeau et al., 2003) and cytokine secretion of NK cells (Andre et al., 2004), rendering NKG2D-mediated pathways important for the elimination of malignant cells (Groh et al., 2001). However, a phenomenon called ‘MICA shedding’ results in the reduction of NKG2D receptor in both, CD8+ tumour-infiltrating T cells and peripheral blood T cells, thereby impairing rather than activating anti-tumour responses (Groh et al., 2002; Kaiser et al., 2007). Herein, MICA proteins are proteolytically cleaved into sMICA upon interaction with the surface of tumour cells. The resulting sMICA:NKG2D complexes are subsequently internalized (due to a missing second signal) and degraded, while immunosuppressive T cells are consequently stimulated (Doubrovina et al., 2003; Groh et al., 2002, 2006; Kaiser et al., 2007). Interestingly, high levels of sMICA in patients undergoing HSCT were correlated with an increased risk of relapse and chronic GvHD (Boukouaci et al., 2009). In addition, tumours have been shown to release microRNAs that downregulate MICA (Kriegeskorte et al., 2005) and to segregate MICA containing exosomes (Ashiru et al., 2010).

3.1.5 MICA and virus infections

In the setting of prolonged virus infection where HLA class I antigen processing and expression are impaired, the interaction between MICA and NKG2D was shown to enhance T-cell cytotoxicity against virally infected cells (Groh et al., 2001). However, viruses could also downregulate MICA gene expression (Stern-Ginossar et al., 2008; Thomas et al., 2008).

3.1.6 129met polymorphism in HSCT

MICA binding to NKG2D is further affected by the SNP rs1051792 that results in an amino acid exchange from valine (val) to methionine (met) at position 129 in the α2 domain of the MICA molecule (Steinle et al., 2001). This SNP alters the binding avidity of MICA to the NKG2D receptor and separates the MICA alleles into weak (129val) and strong (129met) binders (Steinle et al., 2001). The 129met polymorphism was reported to be associated with lower levels of sMICA (Michita et al., 2018) but this effect might be confounded by the presence of atypical membrane anchorage exhibited MICA-A5.1 alleles. However, MICA-129met molecules appear to be expressed at lower levels on the cell membrane due to intracellular retention and to be highly susceptible to membrane shedding (Isernhagen et al., 2015). The SNP rs1051792 has been described to associate with a variety of autoimmune diseases (Amroun et al., 2005; Kirsten et al., 2009; Lopez-Hernandez et al., 2010; Pollock et al., 2013; Raache et al., 2012; Zhao et al., 2011). In the context of HSCT, studies conducted so far reported contradictory results regarding a possible influence of this SNP on the transplantation outcome (Table 2). In one study, patient homozygosity for the 129met variant was associated with higher risk of relapse, whereas patient homozygosity for 129val was resulted in an increased risk of chronic GvHD (Boukouaci et al., 2009). Another study reported that 129met homozygous patients experienced higher risk of acute GvHD yet with a reduced acute GvHD-related mortality (Isernhagen et al., 2015). In three other studies, neither patient nor donor MICA-129 genotype appeared to influence the prevalence of GvHD (acute and chronic, respectively) (Apithy et al., 2018; Askar et al., 2017; Patel et al., 2020). Patient/donor genotype mismatch analyses constitute a further approach to investigate the potential impact of the MICA 129-genotype on HSCT. Herein, MICA-129 genotype matching was associated with approved overall survival after HSCT (Apithy et al., 2018; Fuerst et al., 2016), disease-free survival and lower rates of acute GvHD (Fuerst et al., 2016). However, a large scale study that excluded HLA mismatches found no effect of MICA-129 genotype mismatching on any of the analysed end points (Carapito et al., 2016).

TABLE 2. Overview of published studies on the effect of MICA polymorphism/MICA matching on HSCT outcome

Reference

Year, Tx-country

Patients Study design HLA-matched Variable Overall survival Disease-free survival Relapse aGvHD cGvHD NRM/TRM

Boukouaci

2009, France

211

Related

diverse diagnoses

100%

P 129met/met

P 129val/val

NA

NA

NA

NA

increased

NA

NA

NA

NA

Increased

NA

NA

Isernhagen

2015, Germany

452

Unrelated and related

diverse diagnoses

67.9% and

31.6%

P 129met

P 129met/met

129val/val+ATG

increased

NA

increased

NA

NA

NA

-

NA

NA

decreased

increased

NA

-

NA

NA

-

NA

NA

Askar

2017, USA

713

Unrelated

diverse diagnoses

77,4%

P 129 genotype

D 129 genotype

- - - - - -

Apithy

2018, France

124

Unrelated and related

diverse diagnoses

100%

P 129met

P129val/val

-

NA

NA

NA

NA

- -

NA

NA

Patel,

2019, USA

423

Unrelated and related

diverse diagnoses

100% D 129val/val - - NA - - increased

Boukouaci

2009, France

116

Related

diverse diagnoses

100% P high sMICA NA NA increased NA increased NA

Boukouaci

2009, France

116

Related

diverse diagnoses

100%

P pre-Tx Ab

P post-Tx Ab

NA

NA

NA

NA

NA

increased

NA

NA

decreased

NA

NA

Isernhagen

2015, Germany

452

Unrelated and related

diverse diagnoses

67.9%

31.6%

129 matched increased

NA

NA

NA

NA

NA

Fuerst

2016, Germany

2,172

Unrelated

diverse diagnoses

63.5% 129 mismatched decreased decreased increased (9/10 HLA-matched only) increased - increased

Carapito

2016, France & Netherlands

922

Unrelated

diverse diagnoses

100% 129 mismatched - - - - - -

Kitcharoen

2006, Australia

44

Unrelated

CML

45.5% MICA matched increased NA NA NA NA NA

Parmar

2009, USA

236

Unrelated

diverse diagnoses

73% MICA mismatched NA NA - increased NA NA

Askar

2014, USA

227

Related and cord blood

diverse diagnoses

78% MICA mismatched NA NA NA increased NA NA

Park

2016, Republic of Korea

81

Related and unrelated

diverse diagnoses

69% MICA matched increased - NA - NA NA

Carapito

2016, France & Netherlands

922

Unrelated

diverse diagnoses

100% MICA mismatched - - decreased increased increased increased

Askar

2017, USA

713

Unrelated

diverse diagnoses

77,4% MICA mismatched - - increased - - -

Patel,

2019, USA

423

Unrelated and related

diverse diagnoses

100% MICA mismatched NA NA NA - - NA Note Beneficial effects are highlighted in green; detrimental effects are highlighted in red, respectively. - = no association. Abbreviations: Ab, MICA antibodies; ATG, anti-thymocyte globulin; D, donor; GvHD, graft-versus-host disease; NA, not analysed; NRM, nonrelapse mortality; P, patient; post-Tx, post-transplantation; pre-Tx, pre-transplantation; TRM, treatment-related mortality. 3.1.7 MICA allele mismatch in HSCT

MICA allele mismatch was also analysed thoroughly, and the results found were similarly contradictory (Table 2). Three studies reported an increased incidence of acute GVHD in MICA-allele mismatched cases (Askar et al., 2014; Carapito et al., 2016; Parmar et al., 2009). However, three different studies did not confirm this association (Askar et al., 2017; Park et al., 2016; Patel et al., 2020). Improved overall survival after HSCT was reported in two small scale studies (Kitcharoen et al., 2006; Park et al., 2016), but these findings were not reproduced by two medium scale studies (Askar et al., 2017; Carapito et al., 2016). Only one group demonstrated an increased risk of relapse in the 10/10 HLA-matched subgroup, but not in the 9/10 HLA-matched patients (Askar et al., 2017). A single study reported decreased risk for relapse and increased risk for nonrelapse mortality (Carapito et al., 2016). Hence, regarding the role of MICA-129 genotype mismatching in HSCT, further studies are warranted before definitive conclusions can be drawn.

3.2 MICB 3.2.1 Structure

The major histocompatibility complex (MHC) class I chain-related B (MICB) is located 83 kb centromeric to MICA and exhibits a strong linkage disequilibrium to both MICA and HLA-B, respectively (Bahram et al., 1994). Like HLA and MICA, MICB is a highly polymorphic gene, with 212 alleles reported to date that translate into 37 different proteins (www.ebi.ac.uk/ipd/imgt/hla/stats.html, release 3.42, 2020-10). The amino acid sequence of MICB is highly similar to that of MICA (up to 83% identity) (Bahram et al., 1994). Therefore, it is not surprising that both proteins are similar in terms of certain characteristics. Like MICA, MICB encodes for a single chained glycoprotein that features an HLA-like structure that does not associate with the β2-microglobulin nor does it present peptides (Carapito & Bahram, 2015) but acts as a ligand to the NKG2D receptor on NK cells, γδ T cells and αβ CD8+ T cells (Bauer et al., 1999).

3.2.2 Expression

MICB molecules are expressed constitutively on epithelial cells, fibroblasts, monocytes, dendritic cells and endothelial cells (Carapito & Bahram, 2015; Schrambach et al., 2007). Cell surface expression is upregulated upon cell stress induced by, for example inflammation or neoplastic transformation (Bahram et al., 1994; Groh et al., 1996).

3.2.3 Soluble MICB in HSCT – speculation

Soluble forms of MICB have been detected in the sera of patients in a broad range of tumours or leukaemia, respectively (Groh et al., 1999; Salih et al., 2003, 2006). In addition, MICB can be actively released from epithelial and haematopoietic cells by metalloproteases, thereby reducing the density of NKG2D MICB-ligand on the cell surface. The impairment of MICB binding to NKG2D results in the loss of an activating signal that would have been otherwise strong enough to override the inhibitory signal mediated by HLA-E binding to CD94/NKG2A (Bauer et al., 1999). It has been suggested, therefore, that the proteolytic shedding of MICB may impair anti-tumour and anti-leukaemia immune response, respectively, as part of an immune escape mechanism (Salih et al., 2003, 2006). Interestingly, sMICA and sMICB have been shown to appear either in conjunction or independently from each other (Salih et al., 2006). However, the factors that determine the increase in serum levels of both or only one of these soluble MICs are still unknown.

3.2.4 MICB antibodies in solid organ transplantation

MICB antibodies have been detected in significantly higher levels in the sera of renal transplant patients that suffered graft rejection when compared with functioning-graft patients (Mizutani et al., 2006). MICB molecules are hence potentially important allo-antigens, as they may play a role in solid graft failure.

3.2.5 MICB polymorphisms and CMV

The similarity to MICA ends in the MICB specific capability to bind to the CMV protein UL16 (Cosman et al., 2001). This binding constitutes an immune escape mechanism as it represses the surface expression of unbound MICB molecules that could in turn confer a NKG2D-mediated NK cell activation (Vales-Gomez et al., 2003). The Ile >Met polymorphism at position 98 of MICB*008 (MICB98) results in a decreased binding capacity to UL16 when compared to other MICB alleles (Klumkrathok et al., 2013). In a retrospective study that included more than 900 HLA and MICA matched patients who underwent unrelated HSCT, donor–recipient MICB98 mismatches were associated with an increased incidence of acute and chronic GvHD. In addition, donor–recipient MICB98 mismatches resulted in a higher incidence of CMV infection or reactivation, respectively (Carapito et al., 2020).

Overall, MICB seems to be an interesting candidate when it comes to explain treatment complications like GvHD in fully HLA and MICA matched transplants. Naturally, additional studies have to be conducted in order to investigate this hypothesis.

3.3 Limitations and future recommendations

Within this section, a summary of the limitations of the discussed studies is presented. Additionally, suggestions for studies that aim to overcome these limitations are given (Figure 2).

image

Visual summary of recommendations for future research regarding a possible effect of HLA-E, -F, -G, MICA and MICB, respectively, on HSCT outcome

3.3.1 HLA-E

留言 (0)

沒有登入
gif