WO2019048503A1

WO2019048503A1 - Method for predicting the risk of transplant rejection

Info

Publication number: WO2019048503A1
Application number: PCT/EP2018/073909
Authority: WO
Inventors: Olivier THAUNAT; Alice KOENIG
Original assignee: INSERM (Institut National de la Santé et de la Recherche Médicale); Ecole Normale Superieure De Lyon; Centre National De La Recherche Scientifique - Cnrs -; Universite Claude Bernard Lyon 1; Hospices Civils De Lyon
Priority date: 2017-09-06
Filing date: 2018-09-05
Publication date: 2019-03-14

Abstract

The invention relates to an in vitro method for predicting the risk for transplant rejection in a transplanted subject. The inventors have shown that innate immune effectors NK cells could trigger microvascular inflammation and chronic transplant rejection. NK-mediated rejection was due to the lack of expression by the graft of at least one type of HLA-I ligand for an inhibitory KIR expressed by recipient NK cells Thus the invention relates to an in vitro method for predicting the risk of transplant rejection in a transplanted subject comprising the detection of missing-self activation of NK cells. The inventors also showed that m TOR inhibitors are efficient to prevent missing-self mediated transplant rejection. The invention thus relates to an m Tor inhibitor for use in the prevention or treatment of a transplanted recipient subject at risk of missing-self mediated transplant rejection.

Description

Method for predicting the risk of transplant rejection

FIELD OF THE INVENTION:

The present invention relates to a method for predicting the risk of transplant rejection in recipient subject. The present invention also relates to a method for treating or preventing transplant rejection.

BACKGROUND OF THE INVENTION:

Rejection represents the main cause of solid organ transplants loss and therefore an important unmet medical need in the current context of severe organ shortage.

The current immunological paradigm is that rejection is due to the recognition of alloantigens (mainly the highly polymorphic donor- specific HLA molecules) by recipient's adaptive immune system. Allorecognition, results in the generation of cytotoxic cellular effectors and/or donor specific antibodies (DSA), which drive graft destruction. While the cellular arm of the alloimmune response has progressively come under pharmacological control, current immunosuppressive armamentarium does not completely prevent the generation of DSA (Thaunat et al., 2016). As a result, antibody-mediated rejection (AMR) is now widely recognized as the first cause of allograft loss (Sellares et al., 2012). A few years ago, it has been demonstrated in a murine experimental model of heart transplantation that parental grafts transplanted in Fl recipients developed chronic rejection with graft arteriosclerosis (Uehara et al., 2005). This process was unlikely to be due to adaptive immune system, because in parental => Fl combination the graft do not express any alloantigen). Instead, this experimental model suggested that another type of rejection could lead to vascular lesions and therefore limit the duration of graft function. This model thus clashed with the dogma and the "laws of transplantation" established in the late 60s' claiming that parental into Fl skin transplants was not rejected. This type of rejection had not been demonstrated in the clinical setting. It is therefore needed to be able to predict the risk of transplant rejection in this case, in particular in transplanted patients with no detectable humoral response against the transplant (and notably the transplant endothelial cells) more particularly in transplanted patient who does not present donor specific antibodies. Considering the high morbidity associated with transplant rejection, it also remains highly clinically relevant to propose new adapted treatments for this previously unrecognized condition. SUMMARY OF THE INVENTION:

As mentioned above, the current paradigm is that allograft rejection is driven exclusively by recipient's adaptive immune system. In the present context where cellular immunity, involving mostly CD8 T cells, is controlled by current immunosuppressive drugs, rejection is mostly due to humoral immunity, involving anti donor specific antibodies directed mostly against mismatch donor HLA molecules and leading to microvascular inflammation and to the graft (or transplant) rejection.

The inventors have found that some transplanted patient has microvascular inflammation while having no donor specific antibodies. Surprisingly they discovered that the innate immune effectors NK cells could trigger microvascular inflammation and chronic rejection. NK-mediated rejection was due to the lack of expression by the graft of (at least) one type of HLA-I ligand for an inhibitory KIR expressed by recipient NK cells (a situation known as "missing self"). The present inventors particularly demonstrated that NK-mediated rejection can further occur when HLA I molecules are present at the surface of graft endothelial cells but, because they are allogeneic, cannot bind to KIR inhibitor receptor of NK. Missing self on graft endothelial cells was sufficient to induce the activation of NK cells in the recipient subject, which in turn trigger microvascular inflammation and ultimately lead to transplant rejection (typically chronic transplant rejection).

Therefore the present invention relates to an in vitro method for predicting the risk of transplant rejection in a subject who is the recipient of a transplant from a transplant donor, said method comprising the detection of missing-self activation of NK (natural killer) cells, wherein the detection of missing-self activation of NK cells is indicative of a risk for transplant rejection.

The invention also comprises a method of selecting a HLA compatible donor transplant for a candidate recipient subject comprising:

- obtaining the inhibitory KIR genotype of the candidate recipient subject and the

HLA-I genotype of the transplant donor and the candidate recipient subject;

- comparing the HLA-I genotype and the KIR genotype of the candidate recipient subject et determining the functional KIRs; and - comparing the genotype of the recipient inhibitory functional KIRs of the candidate recipient subject with the donor HLA-I genotype, and

- identifying whether or not a ligand of a functional inhibitory KIR (Killer-cell Immunoglobulin-like Receptor) of the recipient subject is absent in the transplant donor.

The invention further relates to a method for selecting a compatible donor transplant for a candidate recipient subject comprising:

- co-culturing stimulated NK cells from the recipient subject with a population of target cells expressing a defined ligand for a functional inhibitory KIR;

- evaluating the target cells for which a cell death is induced after the co-cultivating step;

- selecting a donor transplant which does not expresses the ligands expressed by the target cells for which a cell-death was induced after the co-cultivating step. The present invention also relates to an inhibitor of mTOR (mammalian target of rapamycin) for use in the prevention of transplant rejection in a subject who is the recipient of a transplant and selected as having at least one recipient inhibitory functional KIR directed against a corresponding HLA-I ligand missing in the donor HLA-I genotype. DETAILED DESCRIPTION OF THE INVENTION:

1. Definitions:

As used herein "transplant rejection" (or graft rejection) encompasses acute transplant rejection and chronic transplant rejection. According to the present invention, the transplant rejection is chronic transplant rejection. In particular, said chronic transplant rejection is associated with the presence of microvascular inflammation in the transplant. Preferably also the transplant rejection is NK cell-mediated rejection, notably missing-self-induced NK- mediated rejection.

Natural killer cells are the third population of lymphocytes defined by the CD3- CD56+ cell surface phenotype and share several features with CD8+ cytolytic T-lymphocytes in their development, morphology, cell surface phenotypes, killing mechanism, and cytokine production. NK cells express both activating and inhibitory receptors that are calibrated to ensure self-tolerance, while exerting early assaults against virus infection and tumor transformation. Having properties of both innate and adaptive immunity, NK cells spontaneously lyse target cells, as well as function as regulatory cells influencing subsequent antigen- specific T-cell and B-cell responses.

The recipient NK cells can recognize and respond against an allograft by three possible mechanisms: missing-self recognition, induced-self recognition, and ADCC (Rajalingam R. Variable interactions of recipient killer cell immunoglobulin-like receptors with self and allogenic human leukocyte antigen class I ligands may influence the outcome of solid organ transplants. Curr Opin Organ Transplant (2008) 13:430-7). Because NK cells circulate in a state that can spontaneously deliver effector function, it is critical that they do not attack surrounding healthy cells. To prevent such detrimental autoreactivity, NK cells express an array of inhibitory receptors recognizing self-HLA class I molecules.

Expression of four distinct HLA class I molecules (HLA-A, -B, -C, and -E) on normal healthy cells provides ligands for the inhibitory receptors of NK cells and, consequently, are resistant to NK cell attack. Down-regulation of HLA class I (HLA-I) expression due to certain viral infections, neoplastic transformations, or absence of relevant HLA class I ligands on the allograft at the setting of allogeneic transplantation, alleviates inhibitory signals, permitting NK cells to eliminate these unhealthy or allogeneic target cells, a phenomenon originally described as the "missing-self hypothesis (see notably Ljunggren HG, Karre K. In search of the "missing self: MHC molecules and NK cell recognition. Immunol Today (1990) 11:237- 44). In addition to the "missing-self mechanism, the expression of ligands for activating receptors on stressed target cell surface might also contribute to NK cell attack, known as "induced-self recognition. The activation receptors can directly recognize stress-induced ligands associated with certain physiological conditions, such as infection, tumor transformation, and transplanted allograft (Raulet DH, Vance RE, McMahon CW. Regulation of the natural killer cell receptor repertoire. Annu Rev Immunol (2001) 19:291-330).

Thus, as used herein, "missing-self-induced NK-mediated rejection" means that transplant rejection is mediated by NK cells which activation is triggered by "missing self.

By "missing self it is herein intended that the inhibitory receptors of NK cells missed their corresponding HLA-I ligand in the target cell, in particular on the transplant cells (and typically endothelial donor cells).

Fourteen distinct KIRs (Killer-cell Immunoglobulin-like Receptors) have been characterized in humans that comprise either 2 or 3 (2D or 3D) Ig-like domains and either a long (L) or short (S) cytoplasmic tail (see table 1 below) (for review see also Rajalingam R. The Impact of HLA Class I-Specific Killer Cell Immunoglobulin-Like Receptors on Antibody-Dependent Natural Killer Cell-Mediated Cytotoxicity and Organ Allograft Rejection. Frontiers in Immunology. 2016;7:585 and Thielens A, Vivier E, Romagne F. NK cell MHC class I specific receptors (KIR): from biology to clinical intervention. Curr Opin Immunol. 2012 Apr;24(2):239-45. Parham P., MHC class I molecules and KIRs in human history, health and survival. Nat Rev Immunol. 2005 Mar;5(3):201-14). Six KIRs are activating types and the remaining KIRs are inhibitory types. Inhibitory KIRS notably comprise KIR2DL1, KIR2DL2, KIR2DL3, KIR3DL1, KIR3DL2, KIR2DL4, KIR2DL5, KIR3DL3. The cytoplasmic tails of the inhibitory KIRs carry an ITIM motif that trigger inhibitory signals upon binding to distinct HLA class I ligands (as detailed in table 1 below). The short-tailed activating KIRs lack ITIM, but carry a positively charged amino acid residue in the transmembrane region that allows the interaction with an adopter chain DAP- 12

KIRs that are expressed on the surface of NK cells recognize allotypic determinants ("KIR ligands") shared by certain HLA class I allele groups. KIR2DL1 recognizes HLA-C alleles with a Lys80 residue (HLA-C2 (Cw2, Cw4, Cw5, Cw6, Cwl5, Cwl7, Cwl8; "group 2' alleles"), KIR2DL2 and KIR2DL3 recognize HLA-C with an Asn80 residue (HLA-Cw3 and related, "group 1" alleles); KIR3DL1 is the receptor for HLA-B alleles sharing the Bw4 supertypic specificity (HLA-Bw4 (B13, B27, B37, B44, B47, B38, B49, B51, B52, B53, B57, B58, B59, B63, B77, B23, B24, B38)). Finally, KIR3DL2 was shown to function as a receptor for HLA-A3/-A11 alleles when bound to Epstein-Barr virus (EBV) peptides (see also below table 1).

On the basis of gene content, KIR haplotypes are broadly classified into two groups. Group A haplotypes have a fixed gene content (KIR3DL3-2DL3-2DP1-2DL1-3DP1- 2DL4-3DL1-2DS4-3DL2) that encode four inhibitory KIRs, 2DL1, 2DL3, 3DL1, and 3DL2, specific for four major HLA class I ligands, C2, CI, Bw4, and A3/A11, respectively, and an activating KIR 2DS4, which is weakly specific for some HLA-C allotypes (CI or C2 epitope), as well as the HLA-A3/11 epitope. In contrast, group B haplotypes are variable both in numbers and combinations of KIR genes, and comprising several genes (2DL2, 2DL5, 2DS1, 2DS2, 2DS3, 2DS5, and 3DS1) that are not part of the A haplotype. Moreover, B haplotypes possess KIRs that have no binding to HLA class I ligands, such as KIR2DL5, 2DS2, 2DS3, and 2DS5. While group A haplotypes contain only KIR2DS4 as an activating gene, group B haplotypes contain up to five activating KIRs - KIR2DS1, 2DS2, 2DS3, 2DS5, and 3DS1. Inheritance of paternal and maternal haplotypes comprising different KIR gene contents generates human diversity in KIR genotypes. For example, homozygotes for group A haplotypes have only seven functional KIR genes, while the heterozygotes for group A and certain group B haplotypes may have all 14 functional KIR genes. All human populations have both group A and B KIR haplotypes, but their incidences vary substantially among populations.

According to the present invention, a recipient inhibitory KIR is considered "functional" (the terms "successfully educated", or "licensed" may also be used interchangeably), if the said recipient expresses its corresponding HLA-I ligand.

The terms "subject" and "patient" are used interchangeably in the present application. A subject of the present invention is a mammal, preferably a human.

In the present application, the terms "recipient", "transplant recipient", "transplanted recipient", "transplanted patient", or "subject recipient" are used interchangeably. In some embodiments a recipient according to the invention is a candidate recipient.

The donor HLA-I corresponds to the HLA-I of the transplant donor. The term "donor" or "transplant donor" are also used interchangeably. The donor or the transplant donor may also be a putative transplant donor.

1. Method for the prognosis or for the diagnosis of transplant rejection and method for selecting a HLA compatible donor transplant for a candidate recipient subject

The method of the invention may be performed before or after transplantation in the subject recipient. Thus, a subject according to the invention may have already received a transplant (transplanted subject) or may be a candidate for transplantation. In some embodiment, the recipient subject may have detectable microvascular inflammation (mvi) associated with the transplant. In some embodiments also, the subject who has received a transplant may also have or not detectable donor specific antibodies (DSA). In one embodiment, a subject who has received a transplant presents microvascular inflammation and does not have detectable DSA. In one other embodiment, a subject who has received a transplant does not present detectable microvascular inflammation and does not have detectable DSA.

The present invention relates to an in vitro method for predicting the risk of transplant rejection in a subject who is the recipient of a transplant from a transplant donor by the detection of missing-self activation of NK (natural killer) cells. According to the invention a situation of "missing self" can be identified by comparison of the genotype of functional inhibitory KIR of the recipient with the HLA-I genotype of the donor.

Indeed, according to the present invention, the identification of the absence in the transplant donor of a ligand for a functional inhibitory Killer-cell Immunoglobulin-like Receptor (KIR) of the recipient subject. In other words identification in the recipient subject of a functional inhibitory KIR missing his corresponding HLA-I ligand in the transplant (or in the putative transplant) is indicative of a risk for transplant rejection in the said recipient subject. In a particular embodiment, the absence of a ligand for a functional inhibitory KIR of the recipient subject is determined on the endothelial cells of the transplant. In other words, it is determined whether endothelial cells of the transplant express said functional inhibitory KIR-ligand.

Thus, for a given subject recipient / transplant donor couple, the lack of expression by the graft of at least one type of HLA-I ligand for a functional inhibitory KIR expressed by recipient NK cells is indicative of a risk for transplant rejection in the recipient.

Accordingly, in a first embodiment, the present invention pertains to an in vitro method for predicting the risk of chronic transplant rejection in a subject who is the recipient of a transplant from a transplant donor, said method comprising the detection of missing-self activation of NK cells, wherein the detection of missing-self activation of NK cells is indicative of a risk for transplant rejection.

The methods according to the present invention allow identifying a missing-self situation between functional inhibitory KIR on NK cells of the recipient and HLA I expressed by graft endothelial cells

The detection of the missing self-activation of NK cells can be achieved by the identification of the absence in the transplant donor of a ligand for a functional inhibitory Killer-cell Immunoglobulin-like Receptor (KIR) of the recipient subject.

The detection of missing-self activation of NK cells can further be achieved by the identification of the absence of expression by the endothelial cells of the transplant, of a ligand for a functional inhibitory KIR of the recipient subject.

The presence of at least one inhibitory functional KIR in the recipient missing its corresponding HLA-I ligands in the transplant donor leads to the activation of NK cells. Thus, according to the invention, the presence of at least one recipient inhibitory KIR vs transplant donor HLA-I ligand mismatch is indicative of a higher risk (or an increase susceptibility) for transplant rejection as compared to a patient wherein no recipient inhibitory KIR vs transplant donor HLA-I ligand mismatch is identified.

Estimation of the risk for transplant rejection may be achieved based on statistical analysis performed on a statistically significant population of transplanted subjects with no recipient inhibitory KIR/ inhibitory KIR Ligand mismatch vs a statistically significant population of transplanted subjects having at least one recipient inhibitory KIR/ inhibitory KIR ligand mismatch.

The risk of presenting a transplant rejection may be further improved in various clinical situations comprising the presence of microvascular inflammation, the pathological history of the patient (notably the ischemia/reperfusion and infectious histories of the patient).

The invention also comprises a method of selecting a HLA compatible donor transplant for a candidate recipient subject, the method comprising:

- comparing the genotype of the recipient inhibitory functional KIRs of the candidate recipient subject with the donor HLA-I genotype, and

Preferably according to the method of selection of the invention, a transplant donor is selected such that no missing self would be created in the recipient subject. In other words, the transplant donor expresses the HLA-I ligands of the inhibitory functional KIRs of the candidate recipient subject.

The invention further comprises a method for selecting a compatible donor transplant for a candidate recipient subject comprising:

- selecting a donor transplant which does not comprises the ligands expressed by the target cells for which a cell death was induced after the co-cultivating step, or, alternatively, selecting a donor transplant comprising the ligands expressed by the target cells for which no cell death occurred after the co-cultivating step.

The "cell death" results from the activation of the stimulated NK cells. Typically the "target cells" used according to the present invention are selected from:

Human arterial endothelial cells isolated from cadaveric organ donors, in particular different primary endothelial cells bearing defined inhibitory KIR ligands (C1, C2 and Bw4); - Human conditionally immortalized glomerular endothelial cells expressing all the defined inhibitory KIR ligands (C1, C2 and Bw4). They can further be genetically engineered to switch off all the different inhibitory KIR ligands;

Tumoral cell lines (721.221 or K562 class-I negative cell lines) transfected with the defined inhibitory KIR ligands (C1, C2 and Bw4).

"A donor transplant which comprises" a defined inhibitory KIR ligand refers to a transplant which comprises endothelial cells expressing said inhibitory KIR ligand. Alternatively, "a donor transplant which does not comprise" a defined inhibitory KIR ligand refers to a transplant in which endothelial cells do not express said inhibitory KIR ligand.

The methods according to the present invention can be used to identify patients at risk for NK-mediated transplant rejection. In particular, the methods according to the invention can be used to quantify this risk so as to decide whether transplantation should be performed.

The above method can further allow quantifying the intensity of missing self-induced NK cell activation so as to determine the individual risk for NK cell-mediated transplant rejection.

Typically, the method for predicting the risk of transplant rejection as well as the method for selecting a HLA compatible donor transplant for a candidate recipient subject of the invention comprises beforehand the steps of:

HLA-I genotype of the transplant donor and of the candidate recipient subject;

- comparing the HLA-I genotype and the KIR genotype of the candidate recipient subject and determining the functional KIRs. Inhibitory KIRs according to the invention are notably as listed above. Preferentially inhibitory KIRs according to the invention are selected from the group comprising KIR2DL1, KIR2DL2, KIR2DL3, KIR2DL5, KIR3DL1, KIR3DL2 and KIR3DL3. Typically, KIR2DL4 is not considered as an inhibitory KIR because it has been reported as having both activating and inhibitory functions (Campbell and Purdy, 2011; Kikuchi-Maki et al., 2003). Most preferably, inhibitory KIRs according to the invention are selected from inhibitory KIRs for which the corresponding ligand is identified. Thus KIRs according to the invention typically comprise KIR2DL1, KIR2DL2, KIR2DL3, KIR3DL1, and KIR3DL2.

In some embodiments, the inhibitory KIR genotype of the candidate recipient subject and the recipient and transplant donor HLA-I genotypes can be obtained by typing the subject recipient HLA-A, HLA-B and HLA-C alleles.

The transplant donor may be categorized in the following inhibitory KIR-ligand groups: HLA-C group 1 alleles (HLA-C1), HLA-C group 2 alleles (HLA-C2), HLA-Bw4 group alleles and HLA-A3/-A11 alleles.

Comparison of the data on HLA-A, HLA-B, HLA-C 1 and HLA-C2 of donors and recipients allows to determine whether HLA-I ligand of KIRs of the recipient are missing. Depending on the embodiment, seven inhibitory KIR/ inhibitory KIR-ligand pairs as defined above, and more preferably five inhibitory KIR/ inhibitory KIR-ligand pairs are analyzed.

In one embodiment of the invention the presence of the HLA-1 Bw4 ligand of KIR3DL1 is assessed in transplant donor. Typically the functionality of KIR3DL1 in the recipient subject is also determined.

The absence of at least HLA-1 Bw4 in a transplant donor is associated with a risk of transplant rejection in particular in recipient subject expressing functional KIR3DL1.

In another embodiment, a putative transplant donor expressing HLA-1 Bw4 is selected for a given candidate recipient expressing KIR3DL1.

Obtaining inhibitory KIR genotype of the candidate recipient subject and the recipient and transplant donor HLA-I genotypes is typically achieved using high resolution typing, such as by PCR-SSO reverse (One Lambda).

Inhibitory KIRs and / or HLA-I typing may be performed on a biological sample from the recipient or the donor. A sample according to the invention can be a body fluid such as for example blood, serum, lymph, or any biological tissue. The biological sample may also be pretreated, for example, by homogenization, extraction, enzymatic and/or chemical treatments as commonly used in the field.

Optionally, the method for predicting the risk of transplant rejection in a subject, or the method for selecting a HLA compatible donor transplant for a candidate recipient subject further comprises a step of determining the proportion of NK cells bearing inhibitory functional KIRs directed against HLA-I ligand missing in donor HLA-I genotype.

The size of the population of NK cells bearing inhibitory functional KIRs directed against HLA-I ligand missing in donor HLA-I genotype allows to quantify the strength of the "missing self". The risk of transplant rejection indeed vary as a function or the proportion of NK cells bearing inhibitory KIRs, directed against HLA-I ligand missing in the donor HLA-I genotype. Typically, a recipient subject having a high proportion of NK cells bearing inhibitory

KIRs, notably functional inhibitory KIRs, directed against HLA-I ligand missing in the donor HLA-I genotype have a higher risk of transplant rejection as compared to a subject having a high proportion of NK cells bearing inhibitory functional KIRs directed against HLA-I ligand missing in the donor HLA-I genotype.

A threshold, below which the risk may be considered as clinically acceptable, may be established based on statistical analysis of the clinical outcome (in terms of transplant rejection) of a population of transplanted patients having various proportion of NK cells bearing inhibitory functional KIRs directed against HLA-I ligands missing in the donor HLA- I genotype.

In one embodiment of a method of selecting a HLA compatible donor transplant for a candidate recipient subject, a transplant donor is selected whose, HLA-I ligands present at least one mismatch with the inhibitory functional KIRs of the candidate recipient, if the proportion of NK cells bearing inhibitory functional KIRs directed against HLA-I ligand missing in donor HLA-I genotype is below a threshold considered as statistically acceptable in terms of clinical outcome.

As a matter of example, determination of the proportion of NK cells bearing inhibitory functional KIRs directed against HLA-I ligand missing in donor HLA-I genotype may be performed as illustrated in the results detailed in the experimental section (see example 2 or 4).

Alternatively, in a method for predicting the risk of transplant rejection in a subject, the detection of missing-self activation of NK cells may be assessed by

- co-cultivating stimulated NK cells from the recipient subject with endothelial cells, preferably primary endothelial cells typically from the transplant donor or with endothelial cells (typically endothelial cell lines) expressing various combination of inhibitory KIR ligands ; and

- detecting the presence of degranulation (using CD 107a as a marker) or expression of chemokines (such as MIP1), as shown in the results.

In a further embodiment, the detection of missing-self activation of NK cells may be assessed by:

- co-cultivating stimulated NK cells from the recipient subject with a population of target cells expressing a defined ligand for a functional inhibitory KIR;

- evaluating the target cells for which a cell death is induced after the co-cultivating step, said cell death being indicative of a missing-self activation of NK cells.

The "target cells" used are as defined above.

Optionally, the method for predicting the risk of transplant rejection in a subject, or the method for selecting a HLA compatible donor transplant for a candidate recipient subject further comprises the in vitro detection of NK cell-mediated rejection.

The detection of NK cell mediated rejection may be achieved as illustrated in the examples. Typically, stimulated NK cells from the recipient subject are co-cultivated with donor endothelial cells (notably primary endothelial cells) expressing or not the inhibitory HLA-I ligands. NK cells are typically stimulated according to the current practice in the field such as with interleukin 2. Detection of the activation of NK cell may be achieved by detection of degranulation (using CD107a as a marker) or expression of chemokines (such as MIP1), as shown in the results.

Detection of NK cell activation is indicative of a high risk for transplant rejection or for a diagnosis of transplant rejection.

Detection of NK cell mediated rejection as described above may be performed routinely, for the monitoring of patient diagnosed with a high risk for transplant rejection, or to whom a prognostic of a risk of transplant rejection was set according to the method as previously defined. In particular, detection of NK cell mediated rejection in patients may be performed in patients, wherein the proportion of NK cells bearing inhibitory KIRs directed against HLA-I ligand missing in donor HLA-I genotype, is elevated (i.e.: typically above a statistically significant threshold).

Taken together the method according to the invention allows the close monitoring of the risk of transplant rejection in a recipient patient such that invasive detection techniques such as biopsy may be avoided or delayed and an adapted curative or preventive treatment may be propose before transplant rejection. Preferably, according to the invention, the transplant is an organ or a tissue. According to the invention, transplant organ is preferably a solid transplant organ, notably selected from heart transplant, lung transplant, kidney transplant, liver transplant, pancreas transplant, intestine transplant, thymus transplant. A tissue transplant encompasses, composite tissue transplant, bones transplant and tendons transplant (both referred to as musculoskeletal grafts), corneae transplant, skin transplant, heart valves transplant, nerves transplant and veins transplant. Most preferably the transplant is a solid organ transplant such as heart, kidney, liver or lung transplant.

Transplant rejection according to the invention is chronic transplant rejection.

2. Methods of prevention and/or treatment of transplant rejection

The methods described above can be used so as to predict NK-mediated rejection, but can also be used in the follow-up of transplanted patients so as to directly diagnose missing self-induced NK cell rejection. The patients thereby identified can be treated with an mTOR inhibitor.

The methods described above can further be used so as to screen patients waiting for an organ transplantation (i.e. on the waiting list for an organ transplant) so as to guide graft allocation. This allows avoiding transplantation of a graft inducing a NK-mediated rejection in the recipient or treating the recipient with mTOR so as to prevent NK-mediated graft rejection.

The present invention thus further relates to a method of preventing or treating transplant rejection, notably NK cell transplant rejection in a transplanted subject in need thereof wherein:

the subject is identified as having missing-self activation of NK, and wherein an effective amount of an inhibitor of the mammalian target of rapamycin (mTor) is administered to said subject.

Preferably the subject is identified as having at least one inhibitory functional KIR directed against a HLA-I ligand missing in donor HLA-I genotype. In other words according to the invention, a subject is selected such that his graft lacks the expression of at least one HLA-I ligand for a functional inhibitory KIR expressed by his NK cells. Preferably the subject is identified as having a risk for transplant rejection, notably transplant rejection mediated by missing self NK activation, by performing the method for predicting the risk of transplant rejection, as previously described.

Preferably the transplant rejection is transplant rejection mediated by missing self NK activation. In a preferred embodiment of the method, the recipient subject has no donor specific antibody (DSA).

Definitions and optional steps of the method for predicting a risk of transplant rejection according to the invention have been described previously and apply to the present methods of treatment and prevention. mTOR, also known as the mechanistic target of rapamycin and FK506-binding protein 12-rapamycin-associated protein 1 (FRAP1), is a kinase that in humans is encoded by the MTOR gene. mTOR belongs to the family of phosphatidylinositol-3-kinase-related kinases (PIKKs). Members in this family are large in size (>2,500 amino acids) and harbor a kinase domain at their C-terminals that shares sequence similarity to phosphatidylinositol-3-kinase (PI3K). Despite having the sequence signature of a lipid kinase, mTOR is a protein kinase that phosphorylates threonine and serine residues in its substrates. In cells mTOR serves as the catalytic subunits of two multi-protein complexes termed as the mTOR complex 1 (mTORC 1 ) and complex 2 (mTORC2) .

mTORCl is a major downstream component of the PI3K/AKT pathway that relays the signals from tumor suppressors PTEN, LKB1 and TSCl/2, and oncoproteins PI3K and AKT. Downstream mTORCl controls cellular biogenesis through regulation of protein synthesis and turnover. It phosphorylates eIF4E binding protein 1 (4EBP1) and ribosomal protein S6 kinase (S6K), two factors involved in translation initiation. Its activity controls protein turnover through repressing autophagy.

mTORC2 is also involved in the PI3K/AKT pathway but its function is independent of mTORCl. It phosphorylates and stimulates AKT activation, and hence plays a critical role in AKT mediated cell survival

The mTor inhibitor according to the invention is an inhibitor of mTor CI.

Inhibitors of mTor suitable for the invention are notably described in Zheng Y, Jiang Y. mTOR Inhibitors at a Glance. Molecular and cellular pharmacology. 2015;7(2): 15-20.

In particular, an mTor inhibitor according to the invention can be selected from rapamycin or one of its analogs termed as rapalogs (sue as Temsirolimus (CCT779), Everolimus (RAD001), or Ridaforolimus (AP23573)), ATP competitive inhibitors, pyrimidine derivatives (such asPP242 and PP30, the morpholino-linked pyrimidine derivatives (such as WAY-600, WYE-687 and WYE354 (37), KU0063794, the triazine derivative OSI-027, AZD8055, AZD2014 or Pink 128).

The mTor inhibitor inhibits mTorCl. Preferably, the inhibitor of mTor is selected from rapamycin and its analogs termed rapalogs. Most preferably, the mTor inhibitor is rapamycin.

The terms "treatment", "treating" or "treat" and the like refer to obtaining a desired pharmacological and/or physiological effect. This effect is preferentially therapeutic in terms of partial or complete stabilization or cure of transplant rejection and/or adverse effects attributable to transplant rejection (notably chronic transplant rejection). Treatment covers any treatment of transplant rejection in a mammal, particularly a human, aimed at inhibiting the transplant rejection symptom(s), (i.e., arresting its development) or relieving the transplant rejection symptoms (i.e., causing regression of the transplant regression or symptoms).

The terms "prevention", "preventing" or "prevent" and the like also refer to obtaining a desired pharmacological and/or physiological prophylactic effect in terms or completely or partially preventing the transplant or a symptom thereof. It covers therefore any preventive treatment of transplant in a mammal, particularly a human, aimed at preventing the transplant rejection or symptom from occurring in a subject, who may be at risk, or predisposed to transplant rejection but has not yet been diagnosed as having it.

The inhibitor of the can be administered by any suitable route, for example, intravenously, intranasally, peritoneally, intramuscularly, orally and other conventional methods.

Said inhibitor according to the invention can be included in a composition. It can be mixed and/or carried with one or more liquid and/or solid pharmaceutically acceptable carriers, ingredients and/or excipients. As used herein, "pharmaceutically acceptable carrier" is intended to include any and all solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like, compatible with pharmaceutical administration. The use of such media and agents for pharmaceutically active substances is well known in the art. Supplementary active compounds can also be incorporated into the composition. The present invention also provides an inhibitor of mTor as defined above, for use in the treatment or the prevention of chronic transplant rejection (in particular NK cell mediated rejection) in a subject who is the recipient of a transplant and who has been selected as having at least one functional recipient KIR directed against HLA-I ligand missing in donor HLA-I genotype.

Preferably the subject has no DSA.

Typically, the subject is identified as having transplant rejection or being at risk of having transplant rejection according to one of the methods as previously described.

In one embodiment, the recipient subject presents at least one functional recipient inhibitory KIR directed against HLA-I ligand missing in donor HLA-I genotype and has a high proportion of NK cells bearing mismatching functional KIRs directed against HLA-I ligands missing in the donor HLA-I genotype. The present invention also relates to the use of an inhibitor of as defined above for the preparation of a medicament for treating or for prevention transplant rejection in a subject who is the recipient of a transplant and selected as having at least one functional recipient inhibitory KIR directed against HLA-I ligand missing in donor HLA-I genotype. FIGURES:

Figure 1: Microvascular inflammation (mvi) in the absence of any donor specific antibodies (DSA) is associated with a kidney graft survival as worst as mvi due to non- complement activating DSA. Renal graft survival curves were compared in the 4 groups: patients diagnosed with mvi and no DSA (mvi+DSA-), patients diagnosed with complement- independent humoral rejection (mvi+DSA+C3d-), patients diagnosed with complement- dependent humoral rejection (mvi+DSA+C3d+) and for control (mvi-DSA-), ns: p>0.05, ***: p<0.001, ****: p<0.0001; Log Rank test.

Figure 2: Missing self triggers an mTORCl pathway in NK cells: Purified NK cells from a healthy donor were co-cultured with HLA-deficient K562 cells.

A. The intensity of the signal corresponding to the phosphorylated form of S6RP was measured at various time points in NK cells (CD56 mask), isolated or in doublet with K562 cells. Data was normalized over baseline; mean + standard deviation. B. The intensity of the signal corresponding to the phosphorylated form of Akt was measured at various time points in NK cells (CD56 mask), isolated or in doublet with K562 cells. Data was normalized over baseline; mean + standard deviation.

Figure 3: Rapamycin prevents missing self-induced NK cell-mediated rejection in vivo: Wild type C57BL/6 mice were transplanted with β2 microglobulin KO heart subjected to 3 hours of cold ischemia. Recipient mice were treated with i) vehicule (control, Ctrl), ii) cyclosporin A (CsA), iii) or rapamycin (Rapa). Heart grafts were harvested 60 days after transplantation for histological analysis. Two independent experiments. A trained pathologist graded the intensity of each elementary lesion on a semi-quantitative scale (score 0-3). Mean + standard deviation, ns: p>0.05; *: p<0.05; **: p<0.01; ****: p<0.0001; One-way Anova.

Figure 4: Percentage of NK cells bearing only one functional inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1).

Figure 5: 5A. Percentage of CD107a+ cells in the different subpopulations of IL2+ NK cells bearing only one inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2). Mismatched licensed cells show significantly more activation as compared to mathed NK cells or mismatched non licensed NK cells. 5B. Percentage of CD107a+ cells in the different subpopulations of IL2+ NK cells bearing a defined inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2).

Figure: 6A. Percentage of ΜΙΡ1β+ cells in the different subpopulations of IL2+ NK cells bearing only one inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2). 6B. Percentage of ΜΙΡ1β+ cells in the different subpopulations of IL2+ NK cells bearing a define inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2).

Figure 7: Impact of immunosuppressants on missing-self-mediated killing. 7A: Graph showing the remaining percentage of B 2 micro KO cells in C57BL/6 mice as function of time. 7B. Graph showing the remaining percentage of B 2 micro KO cells in C57BL/6 mice in the presence of rapamycine.

Figure 8: Can mTOR inhibitors alleviate missing self -induced NK cell mediated rejection in the clinical setting?

A. The CIBERSORT algorithm was used to quantify NK cells in the graft biopsy of

83 renal transplant patients without DSA. Patients were distributed into two groups according to the number of intra-graft NK cells and graft survivals were compared. *: p<0.05; Log Rank test. B. The PBMCs of 24 patients were collected before and 1 month after introduction of everolimus and p-S6RP was quantified in NK cells stimulated (IL-15+) or not with IL-15 (IL- 15-). Representative flow cytometry profiles are shown (left panel). The relative increase of p-S6RP signal in NK cells after stimulation with IL-15 is shown (right panel). Each circle is a patient. Results obtained before (open circles) and after mTOR inhibitor introduction (black circles) were compared. p<0.0001; Wilcoxon Signed Rank test. C. Purified NK cells from 8 healthy volunteers were co-cultured with K562 cells at different effector to target ratios (E:T) for 6 hours in presence (mTOR inh+) or not (mTORinh-) of rapamycin. Viability of target cells was compared. *: p<0.05, **: p<0.01, ***: p<0.001; Mann Whitney test. D. A Heart transplant recipient with missing self-induced NK-mediated rejection was retrospectively identified. The microvascular inflammation (MVI) was blindly graded on a semi-quantitative scale (score 0-3) and the intensity of lesions before and after mTOR inhibitor introduction was compared. Mean ± standard deviation. ***: /?<0.001; Mann Whitney test. E. A renal transplant recipient with missing self-induced NK-mediated rejection was prospectively identified. The microvascular inflammation [glomeruritis (g), score 0-3 and peritubular capillaritis (ptc), score 0-3] was blindly graded on a semi-quantitative scale and the intensity of lesions before and after mTOR inhibitor introduction was compared.

EXAMPLES:

Example 1: Missing self triggers NK-mediated microvascular injuries and chronic rejection of allogeneic kidney transplants

Introduction

Recent lifestyle changes in developed countries and the increased incidence of chronic diseases have set the stage for the accelerated risk of vital organ failure, and its occurrence, which is currently recognised as the leading cause of premature death worldwide, with an estimated cost of -25% of total health expenditures (www.who.int).

The best (often the only) therapeutic option for patients with end-stage vital organ failure is organ transplantation, which restores vital physiologic functions through the surgical substitution of the defective organ by a functioning graft retrieved from a donor. The antigenic determinants that differ between the donor and the recipient (alloantigens), in particular the highly polymorphic molecules from the major histocompatibility complex [MHC, i.e. human leucocyte antigen (HLA) in human], are inevitably recognized by the adaptive immune system of the recipient (Reindl-Shwaighofer et al., 2017), which leads to the failure of the transplanted organ, a process named "rejection". Until the end of the 1970s, the occurrence of acute cellular rejection episodes, i.e. the infiltration and the destruction of the graft by the recipient's cytotoxic T lymphocytes, represented the main obstacle to the success of transplantation. Consequently, considerable efforts were made to develop potent immunosuppressive drugs blocking T cell activation. Introduction of calcineurin inhibitors in the early 1980s led to a dramatic reduction of the incidence of acute cellular rejection and doubled the percentage of functional renal grafts at 1 year post-transplantation (Calne et al., 1978). However, this spectacular progress in the control of T cell alloimmune response barely impacted graft half-life (Lamb et al., 2011) leading to the emergence of the "humoral theory" of chronic rejection (Terasaki et al., 2005). Seminal experimental studies have indeed demonstrated that repeated intravenous administration of alloantibodies were sufficient to trigger the development of typical chronic rejection lesions in allogeneic cardiac grafts transplanted to T and B cell deficient mice (Russel et al., 1994). The theory was later validated in the clinical setting by a large-scale prospective trial showing that renal recipients with circulating alloantibodies directed against donor- specific HLA molecules (anti-HLA DSA) had twice the graft failure rate as those without (Terasaki et al., 2004). First identified in renal transplantation in the 2000s (Poliquen et al., 2015; Sellares et al., 2012; Wiebe et al., 2012) , antibody-mediated rejection (AMR) has since been recognized as the main cause of failure in heart (Loupy et al., 2016), lung (Roux et al., 2016), pancreas (Cantarovich et al., 2011), and vascularized composite (Thaunat et al., 2015) transplantations.

Graft endothelium represents the biological interface between donor alloantigens and host antibodies, which are retained in the recipient's circulation, due to their size (Chen et al., 2018).

Binding of circulating anti-HLA DSA to directly accessible targets expressed by endothelial cells of graft microvasculature sometimes activates the classical complement pathway, which accelerates the rejection process (Loupy et al., 2013; Sicard et al., 2015), but this is not mandatory for the development of chronic humoral rejection lesions (Guidicelli et al., 2016; Hirohashi et al., 2010) .

Engagement of the surface Fc receptors of innate immune effectors by anti-HLA DSA bound to graft microvasculature is indeed sufficient to trigger the release of lytic enzymes that mediate endothelial cell damage. For this reason, the presence of microvascular inflammation in graft biopsy is widely considered as the histological hallmark of AMR (Fishbein et al., 2012; Gupta et al, 2016). The present study challenges this prevalent dogma. Analysing a cohort of 129 renal transplant patients we found that microvascular inflammation in graft biopsy was not mediated by antibodies in almost half of the cases. Instead, genetic analyses suggested that microvascular lesions were due to the direct activation of the recipient's NK cells by graft endothelial cells, which were unable to deliver the inhibitory signals due to the allogeneic nature of their HLA I molecules. The ability of "missing self" to trigger NK cell activation and endothelial cell damage was confirmed in vitro and in murine experimental models in vivo.

Material and Methods

Human study

The study was carried out in accordance with French legislation on biomedical research and the Declaration of Helsinki.

The computer database (DIAMIC) of the Lyon University Hospital pathology department was used to screen all kidney-allograft biopsies (2024 biopsies in 938 patients) performed between September 1st 2004 and September 1st 2012, for microvascular inflammation (MVI+). The biopsies of the 143 patients were systematically reviewed by the same trained pathologist (M. Rabeyrin), who graded the lesions according to Banff classification (Haas et al., 2018). Fourteen patients, whose biopsy analysis did not confirm the presence of MVI lesions (Banff g+ptc score<2) were excluded. Computer-assisted analyses were conducted as described in Sicard et al (Sicard et al., 2017) to quantify NK cells in the patient biopsies.

Clinical data of the 129 patients enrolled in the study was obtained from two independent national registries [Cristal: www.sipg.sante.fr/portail, and Donnees Informatiques Validees en Transplantation (DIVAT); www.divat.fr/] and crosschecked.

Serum samples banked at the time of biopsy (N° of biocollection: AC- 2011-1375 and

#AC-2016-2706) were screened for the presence of anti-HLA donor- specific antibodies (DSA), and, if positive, for the ability of these anti-HLA DSA to bind the complement fraction C3d. These centralized analyses were performed in a blinded fashion with single- antigen flow bead assays according to the manufacturer's instructions (Immucor, Norcross, GA, USA). To rule out the presence of non-HLA donor- specific antibodies, negative sera were tested in endothelial flow cross match assay as described in Le Bas-Bernardet et al (Le Bas-Bernardet et al., 2003).

The steps leading to the distribution of patients into the first 3 groups of patients (MVI+DSA+C3d+, n=40; MVI+DSA+C3d-, n=29; and MVI+DSA-, n=54) are summarized Table 1.

Table 1: Flow chart showing the distribution of the patients in the different groups.

A control cohort, without MVI on graft biopsy, nor circulating DSA (MVI-DSA-, n=75), but matched for the main clinical characteristics of the MVI+DSA- patients, was established from the pool of 938 patients.

Identification of genetically-predicted missing self

Donor and recipient HLA typing were performed by PCR-SSO reverse (One Lambda, Canoga Park, CA, USA). Recipients were genotyped for the 14 KIR genes (2DL1, 2DL2, 2DL3, 2DL4, 2DL5, 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, 3DL1, 3DL2, 3DL3, 3DS1) and 2 pseudogenes (2DP1, 3DP1) by PCR-SSO reverse (KIR SSO Genotyping Test, One Lambda and Lifecodes KIR Genotyping, Immucor).

2DL1, 2DL2, 2DL3, 3DL1, 3DL2 inhibitory KIR receptors educated NK cells only when the recipient expressed their respective HLA ligand: KIR2DL1/C2; KIR2DL2/C1; KIR2DL3/C1; KIR3DL1/Bw4 and KIR3DL2/A*03, *11.

Genetic prediction of missing self was defined as the lack of expression by the graft of the type of HLA molecule able to bind to an educating KIR of the recipient.

Cell preparation and cultures The human erythroleukemia cell line K562, which lacks expression of any MHC molecules, was cultured in RPMI- 1640 (ThermoFisher Scientific, Courtaboeuf, France) complemented with fetal Bovine serum (FBS) 10 % (Dutscher, Brumath, France), L-Glutamine 2 mM (ThermoFisher Scientific), Penicillin 100 U/mL, Streptomycin 100 μΜ and HEPES 25 mM (ThermoFisher Scientific) (hereafter referred to as "complete RPMI").

Primary human arterial endothelial cells were isolated from organ donors (agreement PFS08- 017 from the Agence de la Biomedecine, www . agence-biomedecine . fr) and prospectively stored in the DrVAT biobank (N° of biocollection #02G55). They were cultured in endothelial cell growth medium 2 (Promocell, Heidelberg, Germany) in flasks coated with fibronectin (Promocell) or gelatin 1 % (Sigma, Saint Quentin-Fallavier, France) and used between passages 2 and 7. Peripheral blood mononuclear cells (PBMC) were isolated from the blood of healthy volunteers by Ficoll gradient centrifugation (Eurobio, Courtaboeuf, France). PBMC were cultured overnight at 37 °C in 5% C02 in complete RPMI supplemented with recombinant human IL-2 (R&Dsystems, Minneapolis, MN, USA) or were maintained at 4°C in complete RPMI. NK cells were purified (> 90%) from PBMC by negative selection with magnetic enrichment kits (Stemcell, Grenoble, France).

Flow cytometry

NK cell count

Two hundred microliters of blood were incubated with anti-CD45 (clone 30-F11, 1/400, BioLegend, London, UK), -CD3 (clone SK7, 1/10, BD biosciences, Le Pont de Claix, France) and -CD56 (clone NCAM16.2, 1/10, BD biosciences) antibodies. The samples were then incubated with a Lysing Solution (BD biosciences) to eliminate the red blood cells.

Lymphocyte count was performed with ABX Pentra 60C+ (Horiba, Irvine, CA, USA).

KIR phenotyping

Single cell suspensions of human PBMC were incubated with a fixable viability dye (ThermoFisher Scientific) for 20 minutes at 4°C. After washing, the cells were incubated first with anti-CD19 (clone HIB19, 1/10, BD biosciences), -CD14 (clone M5E2, 1/10, BD biosciences), -CD3 (clone SK7, 1/10, BD biosciences), -CD56 (clone NCAM16.2, 1/10, BD biosciences), -KIR3DL1 (clone DX9, 1/25, BD biosciences), -KIR2DL1/S5 (clone 143211, 1/10, R&Dsystems), and -KIR2DL3 (clone 180701,1/10, R&Dsystems) antibodies for 15 minutes at room temperature and then with anti-KIR2DLl/Sl (clone EB6B, 1/25, Beckman Coulter, Villepinte, France), -KIR2DL2-3/S2 (clone GL183, 1/25, Beckman Coulter), and - KIR3DL1-2 (clone REA168, 1/10, Miltenyi Biotec, Bergisch Gladbach, Germany) antibodies for an additional 15 minutes. The cells were then fixed with paraformaldehyde 2 % (ThermoFisher Scientific) and the sample was stored at 4°C until analysis.

Data collection

Sample acquisitions were made on a LSR FORTESSA or a FACScanto IIR flow cytometer (BD biosciences) and analyses were performed with FlowJo software version 10.0.8rl (Tree Star Inc, Ashland, OR, USA).

Imaging Flow cytometry

Purified human NK cells (105) were mixed with K562 cells at a ratio of 1: 1 in Vbottomed 96-well plates, centrifuged at 100 g for 1 minutes, and incubated 30 min, 1 hour, 2 hours or 3 hours at 37°C at 5% C02. Negative controls were NK cells cultured alone and positive controls were NK cells cultured with IL15 (100 ng/ml, Peprotech).

At indicated time points, the cells were harvested, stained with a fixable viability dye (ThermoFisher Scientific) and then surface stained with anti-CD3 (clone SK7, 1/10, BD biosciences), and -CD56 (clone NCAM16.2, 1/10, BD biosciences) antibodies.

The cells were subsequently fixed, permeabilized (Lysefix/PermlllR fixation/permeabilization kit, BD Biosciences) and stained with anti-Phospho-S6 Ribosomal Protein Ser 235/236 (clone D57.2.2E, 1/50, Cell Signaling Technology, Leiden, The Netherlands) or anti-PAkt S473 (clone M89-61, 1/40, BD biosciences) antibodies.

Sample acquisitions were made on an ImageStream X Mark II (Amnis-EMD Millipore, Darmstadt, Germany) with 40X magnification and analysed with IDEAS software (v6.0).

NK cell activation in vitro

PBMC were cultured overnight in RPMI supplemented with 500 Ul/ml of recombinant human IL-2 (R&Dsystems). Purified NK cells (105 cells) were then mixed with endothelial cells at a ratio of 1: 1 in flat-bottomed 96-well plates, centrifuged at 100 g for 1 minutes, and incubated at 37°C at 5% C02. Anti-CD 107a-FITC (clone H4A3, 5μ1, ThermoFisher Scientific) was added prior the start of the assay. One hour after the beginning of the co- culture, Golgi Stop (BD biosciences) was added to each well. After 4 hours of co-culture, the cells were harvested and surface stained with appropriate antibody combinations to identify KIR subsets. The cells were subsequently fixed and permeabilized (Cytofix/Cytoperm fixation/permeabilization kit, BD Biosciences), stained with anti-MIPls-V450 (clone D21- 1351, 1/40, BD biosciences) antibodies and analysed by flow cytometry. Endothelial cell viability in vitro

PBMC were cultured overnight in RPMI supplemented with 60 Ul/ml of recombinant human IL-2 (R&Dsystems). In each culture well, 10⁴ human primary endothelial cells (either Bw4^" or Bw4⁺) were seeded. After 24h, 10⁵ purified NK cells from KIR3DL1⁺ or KIR3DL1^" donors were added to the culture. When indicated, O^g of anti-KIR3LDl blocking monoclonal antibody (clone DX9, BD biosciences) or an isotype control was added to the cultures.

Endothelial cell viability was monitored every 5 min for 10 h by electrical impedance measurement with an xCELLigence RTCA SP instrument (ACEA Biosciences, San Diego, CA, USA). The cell indexes (CI) were normalized to the reference value (measured just prior to adding NK cells to the culture). Endothelial cell viability in the experimental well was normalized over the control well.

Mice

Wild type C57BL/6 (H-2^b) mice aged 8-15 weeks were purchased from Charles River Laboratories (Saint Germain sur l'Arbresle, France).

C57BL/6 mice in which β2 microglobulin gene has been deleted (hereafter referred as β2 microglobulin KO) lack MHC class I protein expression on the cell surface.

All mice were maintained under exemption of specific pathogenic organisms condition in our animal facility: Plateau de Biologie Experimentale de la Souris (www.sfr- biosciences.fr/plateformes/animal-sciences/AniRA-PBES; Lyon, France).

All animal studies were approved by the local ethical committee for animal research (CECCAPP, www.sfr-biosciences.fr/ethique/experimentation-animale/ceccapp).

Missing-self-induced NK cell-mediated rejection models

Cell transfer model

These experiments were conducted as in Marcais et al (Marcais et al., 2017). Briefly, splenocytes from wild type C57BL/6 or p2-microglobulin KO mice were labelled respectively with carboxyfluorescein diacetate succinimidyl ester (CFSE, 2 μ Μ, ThermoFisher Scientific) and CellTraceViolet (CTV, 2μΜ; ThermoFisher Scientific). Five million cells of each genotype were IV transferred into wild type C57BL/6 recipient mice.

Sixty hours after transfer, splenocytes were isolated and analysed by flow cytometry. The percentage of remaining p2-microglobulin KO cells was calculated using the following formula: % remaining cells = 100 x ( β 2-microglobulin KO cells/wild- type cells) at 60 h (β2- microglobulin KO cells/wild-type cells) in input mix.

Heart transplantation model

Cervical heterotopic heart transplantations were performed as described in Chen et al (Chen et al., 2018).

When indicated, the heart graft was kept at 4°C for 3 hours before transplantation to induce ischemia/reperfusion injuries.

Heart transplants were harvested 60 days after transplantation, fixed in 4% buffered formalin for 24h and embedded in paraffin for haematoxylin and eosin stain and immunohistochemistry. The following primary antibodies were used: anti-mouse CD31 (clone SZ31; 1/50; Dianova, Hamburg, Germany), anti-mouse CD45 (clone 30- Fl l; 1/40, BD biosciences), and anti-Nkp46 (kind gift from Innate Pharma, Marseille, France) to stain, respectively, the endothelial cells, the hematopoietic cells and the NK cells. The sections were revealed by Vectastain ABC HRP Kit (Vector, Peterborough, UK). The amount of labeled cells was semiquantitatively assessed as follows: 0 normal; 1+ minimal or rare foci; 2+ moderate or several foci; 3+ marked or multifocal or diffuse.

Treatment of recipients

When indicated, mice were given intraperitoneal injections of cyclosporine (Sandimmum, Novartis, Rueil-Malmaison, France) 20mg/kg/day or rapamycin (Bio basic, Amherst, NY, USA) 3 mg/kg/day from day-7 to the end of the experiment.

Statistical analyses

For each data set, mean + standard deviation was calculated. For graphical presentation of the same data sets, box plots were generated using Prism software (Version 6.01; GraphPad Software Inc., La Jolla, CA), which present the entire data set distribution. The centre line in the boxes shows the medians; the box limits indicate the 25th and 75th percentiles, the whiskers indicate the 10^th and 90^th percentiles.

Differences between the groups were evaluated by: Mann- Whitney test, fisher's exact test, unpaired i-test, one-way ANOVA followed by a Tukey's post hoc test, or by two-way ANOVA followed by a Sidak's post hoc test, according to the size of the groups and the distribution of the variable. All the tests used were two-sided. The test used for comparison is indicated in the figure legends. Renal graft survivals were compared using the log-rank test. The differences between the groups were considered statistically significant for p< 0.05 and were reported with asterisk symbols (*: /?<0.05; **: /?<0.01; ***: /?<0.001; /?<0.0001). Results

Antibodies are not the sole trigger for graft microvascular inflammation

In kidney transplantation, microvascular inflammation (MVI) is defined as the presence of innate immune effectors in the lumen of peritubular capillaries (peritubular capillaritis, ptc) and/or glomeruli (glomerulitis, g) (Haas et al., 2017). All kidney graft biopsies performed at the Lyon University Hospital between September 2004 and September 2012 (n= 2024 in 938 patients) were retrospectively reviewed and 129 renal recipients with typical graft microvascular inflammation (g+ptc>2) were identified. The clinical characteristics of these patients are presented in Table 2.

Table 2: Clinical characteristics of renal transplant patients

While 69 of these patients had detectable circulating anti-donor HLA antibodies (i.e. "typical AMR"), the remaining ones (60/129, 46,5%) had no detectable circulating anti-donor HLA antibodies (anti-HLA DSA) in highly sensitive solid phase assays.

Previous studies have shown that bona fide humoral rejections can be triggered by non-HLA antibodies directed against either minor histocompatibility alloantigens or autoantigens (Reindl-Schwaighofer et al., 2017; Dragun et al., 2016; Jackson et al., 2015; Sicard et al., 2016; Thaunat et al., 2012). Flow cytometric crossmatch with activated HLA-matched endothelial cells (Le Bas-Bernadet et al., 2003) indeed identified 6 patients (6/129, 4,6%) with non-HLA antiendothelial cells antibodies that could account for graft microvascular inflammation. Based on these results we concluded that in almost half of the cases (54/129, 41,9%; group MVI+DSA-), graft microvascular inflammation is not caused by host humoral response.

Antibody-independent microvascular inflammation impacts graft survival

The binding of high amounts of antibodies to graft endothelium triggers the classical complement pathway (Smith et al., 2008) that is responsible for acute tissue injuries, which dramatically shorten graft survival (Loupy ey al., 2013; Sicard et al., 2015). However, even in the absence of complement activation, donor- specific antibodies can still recruit innate immune effectors and this microvascular inflammation has a detrimental impact on graft survival (Guidicelli et al., 2016).

In line with this data, we observed that among the 69 renal recipients with typical AMR (i.e. circulating anti-HLA DSA and microvascular inflammation on graft biopsy), the 40 patients whose DSA were able to activate the complement cascade in vitro (group MVI+DSA+C3d+, Table 1), had the highest score for C4d deposition in graft biopsy and the worst graft survival (Figure 1).

Interestingly, the 54 patients with antibody-independent microvascular inflammation (group MVI+DSA-) had the same graft survival as the 29 patients with AMR, due to non- complement activating DSA (group MVI+DSA+C3d-) (Figure 1). Graft survivals of these two groups were better than that of MVI+DSA+C3d+ patients but significantly worse than the graft survival of a matched control cohort (group MVI-DSA-) as shown Figure l.This data therefore demonstrates that, regardless of whether it is DSA-dependent or antibody- independent, microvascular inflammation has the same detrimental impact on graft survival. NK cells are present in both types of microvascular inflammation

Antibodies which are unable to activate the complement cascade can still recruit innate immune effectors that can be responsible for antibody-dependent cell mediated cytotoxicity (ADCC), thus leading to chronic humoral rejection (Pouliquen et al., 2015). Seminal experimental studies (Hirohashi et al., 2010), confirmed by subsequent clinical observations (Hidalgo et al., 2010), have demonstrated that among the various subsets of FcY receptor- expressing innate immune effectors, NK cells are crucial for the development of chronic humoral rejection lesions. In line with this data, we observed the presence of NK cells in the graft microcirculation of MVI+DSA+C3d- patients. Interestingly, NK cell infiltration was similar in MVI+DSA- patients, whose microvascular inflammation was not triggered by antibody deposition on the graft endothelium. This data suggests that in chronic rejection, a final common pathway involving NK cells can be triggered either by the humoral arm of the adaptive immune system of the recipients (as widely accepted) or by direct (antibody-independent) activation of innate effectors.

Genetically predicted missing self increases the risk for antibody-independent microvascular inflammation

What could be the stimulus responsible for NK cell recruitment to graft endothelium in the absence of DSA?

NK cell activation is governed by the integration of activating and inhibitory signals. A major class of NK cell receptors involved in this process is killer Ig-like receptors (KIRs). Activating KIRs have a short cytoplasmic tail (KIR-S) and signal through the DAP12 adaptor but their ligands remain poorly defined (Thielens et al., 2012). Inhibitory KIRs have long cytoplasmic tails (KIR-L) containing two ITIMs. Each inhibitory KIR displays two (KIR2DL) or three (KIR3DL) extracellular Ig-domains that confer specificity for HLA-C or HLA-A/B allotypes, respectively (Thielens et al., 2012; Campbell et al., 2011).

KIR locus is highly polymorphic for allele and gene content (Thielens et al., 2012). At the population level, 2 major KIR haplotypes can be defined. Haplotypes A and B share inhibitory KIR-L content but differ strongly in their activating KIR-S content. Haplotype A patients have only one activating KIR (KIR2DS4) whereas those of haplotype B have multiple activating KIRs (Parham et al., 2008). We analyzed the KIR genotype and KIR haplotype of recipients for whom DNA was available and compared patients from MVTDSA- (control group, n=55) and MVI+DSA- (n= 44) groups. No difference was found between the recipients with antibody- independent microvascular inflammation and the controls (Table 2).

Table 2: HLA and KIR genotypes of donors & recipients Because the HLA locus is located on chromosome region 6p21 whereas the KIR locus is on 19ql3.4, HLA and KIR are inherited independently. Consequently, NK cells need to undergo a process of education, in which auto-reactive NK cells (due to the lack of expression of HLA I ligands for inhibitory KIR receptors) are rendered anergic (Shifrin et al., 2014). The HLA I genotype of recipients was therefore analyzed and integrated in the previous analysis. However, even when only educating inhibitory KIRs were considered, no difference was found between recipients of the MVI+DSA and MVI-DSA groups (Table 2).

The "missing self" theory predicts that the role of educated KIR-expressing NK cells is to eliminate HLA-deficient cells that arise during tumoral transformation (Kusunoki et al., 2000) or as a result of MHC I down-regulation that occurs in certain viral infections (Wiertz et al., 1996). Although graft endothelial cells express a normal level of HLA I molecules, their allogeneic nature could theoretically induce a situation in which donor endothelial cells express an HLA I allotype that is unable to interact with an educating inhibitory KIR receptor expressed by recipient NK cells. This situation could trigger a "pseudomissing self" response by recipient's NK cells. To test whether this hypothesis could explain antibody-independent graft microvascular inflammation, we integrated, for each donor/recipient pair, the genetic analyses of i) recipient KIR and ii) recipient HLA-class I (in order to identify educating KIRs). Recipient data was then combined with the donor HLA-class I genotype to identify situations of missing self. In line with our hypothesis, recipients with antibody-independent microvascular inflammation had statistically more genetically predicted missing self (MS) than matched controls.

Allogeneic endothelial cells trigger missing self-induced activation of NK cells in vitro

Upon activation, NK cells kill target cells by directed exocytosis of cytotoxic granules, which can be quantified by the induced cell surface expression of CD 107a (LAMP-1), a transmembrane protein that usually resides in secretory lysosomes. In addition to their cytolytic function, activated NK cells also secrete a variety of soluble factors, including ΜΙΡ-Ιβ. In order to test whether endothelial missing self could trigger activation of NK cells, primary allogeneic human endothelial cells were co-cultured with NK cells purified from the PBMCs of healthy volunteers. After 4 hours of culture, NK cells were recovered and their inhibitory KIR phenotype and activation status (i.e. expression of CD107a and MIP-Ιβ) was assessed at the single cell level by flow cytometry.

Our first analysis focused on the 5 NK cell populations that expressed only one inhibitory KIR. According to the HLA-class I genotypes of the endothelial cells and NK cell donors, 3 distinct situations were identified for each of these NK cell populations: i) absence of missing self (no MS), ii) presence of a missing self for a ligand not expressed by the NK cell donor (uneducated missing self, uneduc MS), or iii) missing self (MS). In line with the clinical data presented above, the 3 groups of NK cells behaved uniformly and did not show any sign of activation after co-culture with the endothelial cells in absence of prior priming. After priming with low dose IL2, NK cells that could specifically detect the absence of expression of a particular HLA class I molecule (MS group) expressed significantly higher levels of both CD107a and ΜΙΡ-β as compared to NK cells that did not express the specific inhibitory KIR (no MS) or that expressed the appropriate inhibitory KIR but were not educated (uneduc MS). This result validates our hypothesis that allogeneic endothelial cells can trigger missing self- induced activation of primed and educated NK cells.

To determine whether some molecular combinations were more prone than others to promote missing self-induced NK cell activation, the previous dataset was re-analyzed considering each inhibitory KIR separately. All inhibitory KIRs were equally able to promote activation of primed and educated NK cells in the absence of their specific HLA class I molecule on the endothelial cells, except for KIR3DL2. This result fits with the conclusion of recent independent reports (Shaw et al., 2012; Yawata et al., 2008) and suggests that KIR3DL2 might not be a functional inhibitory KIR. Taking this notion into account, we re-analyzed our clinical data removing KIR3DL2 from the definition of genetically predicted missing self. Although this simple change did not correct the other limitations of this method (details discussed in the section "Priming and heterogeneity of the NK cell population influence clinical expression of genetically predicted missing self"), it was sufficient to reduce the proportion of recipients with genetically predicted missing self in the group without graft microvascular inflammation [proportion of recipients with genetically predicted missing self in MVI-DSA- group with vs without KIR3DL2: 21/55 (38.2%) vs 15/55 (27.3%)].

A significant proportion of NK cells (25.4 %) express more than one inhibitory KIR on their surface. To determine how these distinct signals contribute to cell activation, we focused the analysis on NK cells that expressed two inhibitory KIRs, one of them being responsible for missing selfinduced activation. Depending on the type of the second inhibitory KIR and the HLA I genotypes of the endothelial cells and the NK cell donor, 3 situations were identified: i) missing self + matched (MS+M), ii) missing self + uneducated missing self (MS+uneduc MS), or iii) missing self + missing self (2MS). Activation status of the NK cells of these 3 groups after co-culture with allogeneic endothelial cells was compared to that of NK cells that express only one functional inhibitory KIR. The level of expression of both CD107a and ΜΙΡ-1β was increased in 2MS, and decreased in MS+uneduc MS and the MS+M group. These results demonstrate that the signals generated by each functional inhibitory KIR expressed on the surface are integrated by NK cells and modulate missing self-induced activation.

Missing-self-induced activation of NK cells has a deleterious impact on endothelial cells

Having demonstrated that allogeneic endothelial cells could trigger missing self-induced activation of NK cells, we aimed at determining its impact on endothelial cells. To address this issue, the integrity of adherent endothelial cells was monitored by real time impedance measurement in the co-culture model described above. In a first set of experiments, we compared the survival of the same primary endothelial cells exposed to NK cells from two distinct allogeneic donors: one donor with missing self and the other without (negative control). The experiment, reproduced with 6 different pairs, demonstrated that endothelial cell survival was consistently reduced when co-cultured with NK cells expressing inhibitory KIR unable to interact with the appropriate HLA I molecule on endothelial cells.

To rule out the possibility that the differences observed in the first model were influenced by inter-individual heterogeneity of NK cell populations between donors, we developed a second model, in which the endothelial cells were co-cultured with NK cells from the same matched allogeneic donor with anti-KIR3DLl blocking mAb or an isotype control mAb. In line with previous results, co-cultures with anti-KIR3DLl blocking mAb induced an "artificial" missing self, which significantly lowered endothelial cell survival.

Collectively, this in vitro data supports the notion that missing self-induced activation of NK cells has a deleterious impact on graft vasculature.

Missing self triggers NK cell-mediated rejection in vivo

We next investigated the impact of missing self-induced activation of NK cells in two in vivo models. In the first model, fluorescently labelled splenocytes purified from wild type C57B/L6 (controls) and p2-microglobulin KO mice (lacking MHC I) from the same genetic background were IV co-injected to wild type C57B/L6 mice. After 2.5 days, both types of splenocytes were enumerated in the spleen of recipient mice by flow cytometry. In line with in vitro data, missing self triggered the specific destruction of the cellular targets that lacked MHC I expression. The role of NK cells in missing self-induced cellular destruction was demonstrated by the persistence of p2-micro globulin KO splenocytes in recipients, whose NK cells were depleted by anti-NKl. l mAb before splenocytes transfer. To validate the existence of missing self-induced NK-mediated rejection in the context of transplantation, we adapted the heterotopic heart transplantation model. Heart grafts, harvested in wild type C57B/L6 (controls) or p2-microglobulin KO mice, were transplanted to wild type C57B/L6. As observed in the clinic, the mere absence of MHC I molecules on the graft endothelium was insufficient to promote the development of histological lesions. However, priming of the recipients' NK cells induced by mild ischemia/reperfusion injuries resulted in the appearance of microvascular inflammation, specifically in p2-microglobulin KO heart transplants.

Similar results were obtained when priming of NK cells was performed with Poly (I:C), used as a surrogate for viral infection. Graft microvascular inflammation in this model was very similar to that observed in MVI+DSA- patients: circulating CD45+ immune cells were found to adhere to CD31+ turgid capillary endothelial cells, in the absence of complement fraction C4d deposition. The central role of NK cells in this type of rejection was demonstrated by the complete disappearance of lesions in β2-microglobulin KO heart grafts transplanted to recipients, whose NK cells were depleted by anti-NKl.1 mAb.

Missing self triggers mTORCl pathway in NK cells

To gain insights into the molecular mechanisms involved in missing self-induced NK cell activation, human NK cells were purified from PBMCs of healthy volunteers and co-cultured with K562 cells, an MHC class I-deficient human cell line. Based on previous works from our group, which have reported the critical importance of this pathway for NK cell activation (Marcais et al., 2017 and 2014), the analysis was focused on mTOR pathway. The phosphorylation status of S6 Ribosomal Protein (S6RP) and protein kinase B (AKT), located downstream from mTORCl and mTORC2 complexes respectively, was longitudinally assessed in NK cells using imaging flow cytometry (Data not shown). While isolated NK cells showed only a modest increase of pS6RP, the mTORCl pathway was strongly activated in NK cells that had formed doublets with K562 targets. In contrast, no significant change was observed regarding the phosphorylation status of AKT in NK cells, which suggests that mTORC2 does not play a role in missing self-induced NK cell activation (Figure 2).

Rapamycin blocks mTORCl in vivo and prevents missing self-induced NK-mediated rejection

Calcineurin inhibitors are currently the cornerstone of therapeutic immunosuppression in solid organ transplantation (Ekberg et al., 2007; Thaunat et al., 2016). Rapamycin, an allosteric inhibitor of the mTORCl complex, was approved for immunosuppression as an alternative to calcineurin inhibitors in the early 2000s (Li et al., 2014). Based on the molecular data presented above, we hypothesised that rapamycin might have potent therapeutic effects against missing self-induced NK-mediated rejection. To test this theory we compared the effects of rapamycin and cyclosporin in the two in vivo models presented above. As expected, rapamycin, but not cyclosporin, reduced the activation of the mTORCl pathway in NK cells exposed to missing self in vivo. The blockade of mTORCl with rapamycin correlated with significantly improved survival of p2-microglobulin KO cellular targets as compared with controls and cyclosporin treated animals. The beneficial effect of rapamycin was also observed in the heterotopic heart transplantation model. Indeed, while cyclosporin A-treated animals developed the same microvascular inflammation as untreated controls, recipient mice treated with rapamycin showed significantly less endothelial turgidity and inflammatory effectors in heart graft capillaries (Figure 3).

Our data therefore validate the idea that rapamycin may protect transplant recipients against missing self-induced NK-mediated rejection.

Discussion

In the present study, we demonstrated that the allogeneic nature of graft endothelial cells sometimes creates missing self, a situation that can be sensed by primed NK cells in the recipient's circulation. Missing self-induced NK cell activation promotes the development of graft microvascular inflammation that has the exact same detrimental impact on organ survival as non-complement activating anti-HLA DSA, the primary cause of late transplant loss However, while there is currently no efficient therapy against antibody-mediated chronic vascular rejection, our study established that missing self-induced activation of NK cells is dependent upon the mTORCl pathway that can be blocked by rapamycin. Preclinical studies using experimental murine models confirmed the efficiency of rapamycin to prevent the development of histological lesions.

We believe that this data can have several levels of significance. Firstly, clinicians in charge of transplant patients are frequently confronted with microvascular inflammation lesions on graft biopsy. As an illustration, the prevalence of Banff score g+cpt>2 lesions was estimated to be as high as 13.8% in our cohort of renal transplant patients, albeit we do not perform HLA incompatible transplantations (i.e. transplantation in the presence of preformed anti- HLA DSA) in our centre. Until now, microvascular inflammation lesions have been considered as the hallmark of antibody-mediated rejection (AMR). Induced by the growing importance of AMR, now recognized as the primary cause of transplant failure, the consensus, established in 2013 by the Banff group, stated that patients with MVI should be diagnosed with AMR, despite the possible absence of circulating anti-HLA DSA and of positive C4d staining on the graft biopsy. The rationale for this decision was based on the fact that available assays only detect anti-HLA antibodies, whereas graft endothelial cells express other minor histocompatibility antigens or autoantigens (Reindl-Schwaighofer et al., 2017; Le Bas-Bernadet et al., 2003; Dragun et al., 2016; Jackson et al., 2015; Joosten et al., 2002) that can be targeted by the recipient's humoral immune response. Our data suggest that such anti- endothelial cell antibodies exist but account for <10 of isolated MVI cases. Instead, we propose that most isolated (DSA-) MVI are the result of a previously unrecognized type of rejection, due to the "direct" activation of the recipient's NK cells by missing self.

Although missing self-induced and non-complement-activating anti-HLA antibody-mediated rejections have the same detrimental impact on graft survival, it is crucially important to differentiate these two conditions. Patients with missing self-induced rejection will indeed not respond to the costly and tedious treatment of AMR, which associates plasmapheresis with high dose intravenous immunoglobulins (Pouliquen et al., 2015). The mixture of authentic cases of AMR with previously unrecognized cases of missing self-induced rejection might explain the high heterogeneity in response to treatment. Furthermore, our results demonstrated that rapamycin efficiently prevents the development of histological lesions due to missing self-induced NK cell activation in a murine experimental model of transplantation. Of note, on the basis of conflicting clinical reports suggesting that mTOR inhibitors might be less potent to prevent DSA generation (Thaunat et al., 2016; Grimbert et al., 2017; Kamar et al., 2013; Liefedt et al., 2012), transplanted patients with graft microvascular inflammation (and wrongly diagnosed with AMR) are often switched from an mTOR inhibitor-based to calcineurin-inhibitor (CNI) based maintenance regimen. This is likely detrimental to graft survival because our data shows that CNI have no impact on missing self-induced rejections. In conclusion, this study identifies a new type of chronic rejection, whose pathophysiology is independent of recipient's adaptive immune system. Missing self-induced NK cell-mediated chronic vascular rejection is as prevalent as AMR and has the same detrimental impact on organ survival. However, while there is currently no efficient therapy against chronic AMR, rapamycin, a commercially available mTOR inhibitor, has shown promising efficiency to prevent the development of histological lesions in a preclinical murine model of missing self- induced NK cell-mediated chronic vascular rejection. Example 2: In vitro modelisation of missing-self mediated NK activation

Materiel and Methods

At day 0, whole blood samples were collected from healthy donors by venipuncture into Acide Citrate Dextrose-containing vials. Peripheral blood mononuclear cells (PBMC) were isolated by Ficoll gradient centrifugation (Eurobio). PBMCs were cultured overnight at 37°C in 5% C02 in complete culture medium (RPMI 1640 containing glutamine and supplemented with 10 % FBS, hepes and penicillin-streptomycin) supplemented with 500 Ul/ml recombinant human IL-2 (R&Dsystems) or were maintained at 4°C in complete culture medium.

Simultaneously, endothelial cells were seeded (100 000 in each well) in wells of a flat bottom 96- well plate coated with gelatin 1% (Sigma). Endothelial cells were cultured overnight at 37 °C in 5% C02 in endothelial cell growth culture 2 medium (Promocell).

A day 1, NK cells were isolated using stem cell magnetic kit according to the manufacturer's instructions. Then, purified NK cells were resuspended at 0.5 million/ml in complete RPMI.

One hundred thousand NK cells were added in each well containing endothelial cells after removing endothelial cell culture medium. Five microliter of anti-human CD 107a FITC (eBIOH4A3, ebioscience) was added in each well at the beginning of the co-culture. One hour after the beginning of the co-culture, golgi stop (DB biosciences) was added in each well. Then cells were co-cultured for 3 hours.

After the co-culture, cells were detached with trypsin and recovered in V bottom- 96- well plates.

Then cells were stained 20 min at room temperature in 50 μl of the following antibodies diluted in PBSlx:

CD3 APC-H7 (SK7) BD biosciences 1/25e

CD56 PE CF594 (NCAM16.2) BD biosciences 1/25e

KIR2DL1/S5 AF 700 (143211) R&Dsystems 1/10e

KIR2DL3 APC (180701) R&Dsystems 1/10e

KIR3DL1 BV 711 (DX9) BD bisociences 1/25e

CD 107a FITC (eBIOH4A3) ebiosciences 1/50e

Fixable viability dye eFluor 506 ebiosciences 1/1000e Without washing, the following antibodies were added in 50 μΐ in PBSlx for an incubation of 15 minutes at room temperature:

KIR2DL1/S1 PE CY7 (EB6B) Beckman Coulter 1/25e

KIR2DL2-3/S2 PE CY5.5 (GL183) Beckman Coulter 1/25e

KIR3DL1-2 PE (REA168) Miltenyi Biotec 1/10e

Then cells were washed twice with PBSlx. After this, cells were fixed with 75 μΐ of Cytofix/Cytoperm® fixation/permeabilization kit (BD Biosciences) 20 min at 4°C. After two washing with Permwash, cells were stained 30 min at 4°C in 100 μΐ of the following antibody diluted in Permwash:

MIP1B V450 (D21-1351) BD biosciences 1/40e

After one washing with PBSlx, cells were resuspended in PBSlx. Then samples acquisitions were made on a BD FORTESSA IV flow cytometer (BD Biosciences). Data were analysed with FlowJo software (Tree Star).

Results:

NK cells from healthy donors previously stimulated (IL2+) or not (IL2-) with IL2 were co-cultured 4 hours with different primary endothelial cell lines bearing different HLA class I ligands of inhibitory KIRs.

NK cell activation is assessed by the expression of the degranulation marker CD107a or the chemokine ΜΙΡΙβ.

The percentage of NK cells bearing only one inhibitory KIR expressing the degranulation marker CD107a or the chemokine ΜΙΡΙβ was assessed by flow cytometry in the following subpopulations :

NK cells with an iKIR having its ligand present on endothelial cells (Matched iKIR (M)),

non licensed NK cells with a iKIR not having its ligand on endothelial cells (Mismatched non licensed iKIR (MM L-)), and

- licensed NK cells with an iKIR not having its ligand on endothelial cells

(Mismatched licensed iKIR (MM L+)).

After 4 hours of co culture the percentage of NK cells expressing CD107a or MIP1B was calculated for the following NK cells populations (KIR repertoire staining and analysis in NK cell was performed as described in Beziat V, Liu LL, Malmberg J-A, et al. NK cell responses to cytomegalovirus infection lead to stable imprints in the human KIR repertoire and involve activating KIRs. Blood. 2013;121(14):2678-2688).:

NK cells bearing one inhibitory KIR which had its ligands on the endothelial cell line (matched NK cells);

- NK cells bearing one one non-functional inhibitory KIR which doesn't have its ligand on endothelial cells (mismatched non licensed NK cells); and

NK cells with one functional inhibitory KIR which doesn't have its ligand on endothelial cells (mismatched licensed NK cells).

In the absence of NK cell stimulation with IL2 before the co-culture, none of the 3 subpopulations of NK cells expressed either CD 107 A or MIPIB. However, when they were stimulated with IL2 before the co-culture, the percentage of mismatched licensed NK cells positive for the degranulation marker CD107a and the chemokine MIPIB was significantly higher than for matched NK cells and mismatched non licensed NK cells. These results were found for the 3 inhibitory KIRs which can be licensed: KIR2DL1, KIR2DL2-3, KIR3DL1. The results confirm that NK cells can sense missing-self on primary endothelial cells.

These results also suggest that NK cells need to be primed to sense missing-self on endothelial cells, as without IL2 nothing occurs.

Figure 4 shows the percentage of NK cells bearing only one functional inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1).

Figure 5 A shows the percentage of CD107a+ cells in the different subpopulations of

IL2+ NK cells bearing only one inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2). Mismatched licensed cells show significantly more activation as compared to mathed NK cells or mismatched non licensed NK cells.

Figure 5B shows the percentage of CD107a+ cells in the different subpopulations of IL2+ NK cells bearing a define inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2).

Figure 6 A shows the percentage of ΜΙΡ1β+ cells in the different subpopulations of IL2+ NK cells bearing only one inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2). Figure 6B shows the percentage of ΜΙΡ1β+ cells in the different subpopulations of IL2+ NK cells bearing a define inhibitory KIR (KIR2DL1 or LIR2DL2 and/or 3, KIR3DL1, KIR3DL2).

Example 3: Identification of immunosuppressant to be used for transplant rejection mediated by missing-self activated NK cells a) In vitro model:

mTOR pathway was selected as a target of interest as mTOR was known to be involved in NK cell activation. Furthermore inhibitors of mTOR, such as rapamycin, are already available in clinic. The same experimental setting as previously described (see example 2) was used.

The results show that mTOR efficiently prevents NK cell mediated transplant rejection. b) In vivo cellular model of missing self

Wild type B6 mice were, injecting with splenocytes coming from a B2 micro KO mouse (lacking class I molecules on their surface) and therefore sensitive to the NK dependent lysis by missing self.

To control that the lysis of the cells is not due to another phenomenon, splenocytes from a wild-type B6 mouse were injected at the same time.

After the injection, the recipient mice were killed at different time points and their spleen were recovered. The splenocytes were analyzed by flow cytometry and the % of remaining B2 micro KO cells compared to that injected at day 0 was calculated.

With the time, B 2 micro KO cells are progressively killed and disappear around day 7 (see figure 7A).

In mice treated with an inhibitor of mTORCl, such as rapamycine, survival of B2 micro KO cells is drastically increased which confirm the potential efficacy of this drug in the treatment of transplant rejection mediated by missing-self-activated NK cells.

Example 4: Predicting missing self NK-mediated microvascular injuries

At day 0:

Whole blood samples are collected from patients by venipuncture into heparinate-containing vials. Peripheral blood mononuclear cells (PBMC) are isolated by Ficoll gradient centrifugation (Eurobio). PBMCs are then cultured overnight at 37°C in 5% C02

in complete culture medium (RPMI 1640 containing glutamine and supplemented with 10 % Fetal Bovine Serum (FBS), hepes and penicillin- streptomycin) supplemented with 500 Ul/ml recombinant human IL-2 (R&Dsystems) or maintained at 4°C in complete culture medium. The following types of target cells can be used:

Human arterial endothelial cells isolated from cadaveric organ donors, in particular different primary endothelial cells bearing different inhibitory KIR ligands (CI, C2 and Bw4); - Human conditionally immortalized glomerular endothelial cells expressing all the different inhibitory KIR ligands (C1, C2 and Bw4). They can further be genetically engineered to switch off all the different inhibitory KIR ligands;

Tumoral cell lines (721.221 or K562 class-I negative cell lines) transfected with the different inhibitory KIR ligands (C1, C2 and Bw4).

The target cells are then transfected with the nanoluciferase.

If the target cells are endothelial cells, they are seeded (100 000 in each well) in wells of a flat bottom 96-well plate coated with gelatin 1% (Sigma) and cultured overnight at 37°C in 5% C02 in endothelial cell growth culture 2 medium (Promocell).

If the target cells are tumoral cell lines, they are seeded (100 000 in each well) in wells of a flat bottom 96-well plate and cultured overnight at 37 °C in 5% C02 in complete culture medium. A day l,

NK cells are isolated using stemcell magnetic kit according to the manufacturer's instructions. Then, purified NK cells are resuspended at 0.5 million/ml in complete RPMI.

One hundred thousand NK cells are then added in each well containing endothelial cells after removing endothelial cell culture medium. Five microliter of anti-human CD 107a FITC (eBIOH4A3, ebioscience) are added in each well at the beginning of the co-culture. One hour after the beginning of the co-culture, golgi stop (DB biosciences) is added in each well. The cells are then co-cultured for 3 hours.

After the co-culture,

· First, the supernatant is recovered for each well in V bottom- 96-well plates. The plates are centrifuged so as to have a clean supernatant without cells. The substrate of the luciferase is then added in each well and after a determined time, the luminescence of each well is tested to evaluate the target cell death. • Cells separated from the supernatant are kept in V bottom- 96-well plates and mixed with the cells remaining on the bottom of the flat bottom 96-well plate detached with trypsin. The cells are then stained 20 min at room temperature in 50 μΐ of the following antibodies diluted in PBSlx:

CD3 APC-H7 (SK7) BD biosciences l/25e

CD56 PE CF594 (NCAM16.2) BD biosciences l/25e

KIR2DL1/S5 AF 700 (143211) R&Dsystems 1/lOe

KIR2DL3 APC (180701) R&Dsystems 1/lOe

KIR3DL1 BV 711 (DX9) BD bisociences l/25e

CD 107a FITC (eBIOH4A3) ebiosciences l/50e

Fixable viability dye eFluor 506 ebiosciences 1/lOOOe

Without washing, the following antibodies are added in 50 μΐ in PBSlx for an incubation of 15 minutes at room temperature:

KIR2DL1/S 1 PE CY7 (EB6B) Beckman Coulter l/25e

KIR2DL2-3/S2 PE CY5.5 (GL183) Beckman Coulter l/25e

KIR3DL1-2 PE (REA168) Miltenyi Biotec 1/lOe

The cells are then washed twice with PBSlx. After this, cells are fixed with 75 μΐ of Cytofix/Cytoperm® fixation/permeabilization kit (BD Biosciences) 20 min at 4°C. After two washing with Permwash, cells are stained 30 min at 4°C in 100 μΐ of the following antibody diluted in Permwash:

MIP1B V450 (D21-1351) BD biosciences l/40e After one washing with PBSlx, cells are resuspended in PBSlx. Then samples acquisitions are made on a BD FORTESSA IV flow cytometer (BD Biosciences). Data are analyzed with FlowJo software (Tree Star).

Example 5 :

Can mTOR inhibitors alleviate missing self-induced NK cell-mediated rejection in the clinical setting?

Material and Methods

Independent validation cohort Eighty-three patients that received an isolated kidney allograft at the renal transplantation center of the KU Leuven were enrolled in the validation cohort. Inclusion criteria were: >18 years of age, absence of detectable DSA in the circulation by solid phase assay, available data from transcriptomic analysis of renal graft biopsy.

Transcriptomic analysis of renal graft biopsies

For each renal allograft biopsy included in this study, at least half a core was immediately stored on Allprotect Tissue Reagent® (Qiagen Benelux BV, Venlo, The Netherlands), and after incubation at 4°C for at least 24 hours and maximum 72 hours, stored at -20°C. RNA extraction was performed using the Allprep DNA/RNA/miRNA Universal Kit® (Qiagen Benelux BV, Venlo, The Netherlands) on a QIAcube instrument (Qiagen Benelux BV, Venlo, The Netherlands). The quantity (absorbance at 260nm) and purity (ratio of the absorbance at 230, 260 and 280nm) of the RNA isolated from the biopsies were measured using the NanoDrop ND-1000™ spectrophotometer (Thermo Scientific™, Life Technologies Europe BV, Ghent, Belgium). RNA integrity was evaluated using the Eukaryote nano/pico RNA Kit® (Agilent Technologies Belgium NV, Diegem, Belgium) on the Bioanalyzer 2100 instrument™ (Agilent Technologies Belgium NV, Diegem, Belgium). Samples were stored at -80°C until further analysis.

RNA extracted from the biopsy samples was first amplified and biotinylated to complementary RNA (cRNA) using the GeneChip® 3' IVT PLUS Reagent Kit (Affymetrix Inc., High Wycombe HP10 0HH, UK) and subsequently hybridized onto Affymetrix GeneChip Human Genome U133 Plus 2.0 Arrays (Affymetrix Inc., High Wycombe HP10 0HH, UK), which covers over 54k transcripts, according to the manufacturer's instructions. The arrays were scanned using the GeneChip® Scanner 3000 7G System (Affymetrix Inc., High Wycombe HP 10 0HH, UK), and image files were generated using the GeneChip® Command Console® Software (AGCC). Finally, Robust Multichip Average (RMA) background correction and normalization was performed using the Affymetrix Expression Console Software, and expression values were log2 scaled. Eighty-three biopsies survived pre-hybridization quality control checks, and were analyzed.

CIBERSORT algorithm

CIBERSORT (Cell type Identification By Estimating Relative Subsets Of known RNA Transcripts) is a recently introduced computational method allowing the exploration of the heterogeneity of infiltrating immune cells in complex solid tissues (Newman et ah, Nature methods 12, 453-457 (2015). This deconvolution algorithm calculates the relative fraction of different phenotypes of human hematopoietic cells in tissue samples, based on the expression of 547 genes. To obtain an estimation of absolute intragraft NK-cell count, we adjusted the CIBERSORT data for the expression of pan-leukocyte marker CD45/PTPRC.

Identification of informative cases of

Heart transplantation

The databases of the pathology department of Georges Pompidou University Hospital and the HLA Laboratory of Saint Louis Hospital (Paris, France) were cross-interrogated to retrospectively identify heart transplant recipients with the following characteristics: i) presence of MVI on graft biopsies, ii) absence of circulating anti-HLA DSA, iii) delayed introduction of an mTOR inhibitor in maintenance regimen. Genetic analysis confirmed the existence of a missing self: recipient's NK cells expressed functional inhibitory KIR2DL3 receptors while donor lack the expression of HLA-C1 molecules.

Heart graft biopsies of the patient were fixed in formol and paraffin embedded sections were stained by routine methods. Microvascular inflammation (MVI) was blindly graded by a trained pathologist on a semi-quantitative scale (score 0-3). Phosphorylated-S6RP staining was performed by immunohistochemistry on frozen sections using an anti-human p-S6RP clonal antibody (clone D68F6, produced by cell Signaling Technology, Leiden, The Netherlands).

Renal transplantation

A renal transplant patient from Lyon University Hospital was biopsied 3 months' post- transplantation systematically. Analysis of renal biopsy revealed significant MVI lesions despite the absence of detectable DSA in the circulation. Negative endothelial cross-match ruled out the diagnosis of AMR due to non-HLA DSA. Genetic analyses identified 2 missing self: recipient's NK cells expressed functional inhibitory KIR2DL1 and KIR3DL1 receptors while donor lack the expression of HLA-C2 and Bw4 molecules. This led to the diagnosis of missing self-induced NK-mediated rejection. The antimetabolite was replaced by an mTOR inhibitor (everolimus) and a new biopsy was performed 3 months after the modification of maintenance immunosuppression. Renal graft biopsies were fixed acetic acid-formol-absolute alcohol and paraffin embedded sections were stained by routine methods. MVI lesions were blindly graded by a trained pathologist according to the Banff classification updated in 2014 [glomeruritis (g), score 0-3 and peritubular capillaritis (ptc), score 0-3].

IL15-induced mTORC1 activation in NK cells

PBMCs of 24 patients with a breast cancer were collected before and one month after the introduction of everolimus. PBMCs were cultured for 1 hour in complete RPMI. When indicated, 100 ng/ml of IL-15 was added to the cultures. After 1 hour, the cells were harvested and surface stained with appropriate antibody combinations. The cells were subsequently fixed and permeabilized (Cytofix/Cytoperm fixation/permeabilization kit, BD Biosciences), stained with anti-Phospho-S6 Ribosomal Protein Ser 235/236 (clone D57.2.2E, 1/50, Cell Signaling Technology, Leiden, The Netherlands) antibody and analysed by flow cytometry. K562 cell viability in vitro

NK cells were co-cultured with 2500 K562 cells transfected with NanoLuc® luciferase at different effector to target ratio. When indicated, 25 nM of rapamycin was added to the cultures. After 6 hours of co-culture, 50 μΐ of supernatant of each well was collected and Nano-Glo® Luciferase Substrate (Promega, Madison, WI, USA) was added. K562 cell viability was assessed by measurement of luminescence for each well with an Infinite® 200 PRO instrument (TEC AN, Mannedorf, Switzerland).

Results

The notion that NK cells can promote transplant destruction independently of DSA was validated in an independent clinical cohort. A recently reported computational method named CIBERSORT (Cell type Identification By Estimating Relative Subsets Of known RNA Transcripts, (Newman et ah, Nature methods 12, 453-457 (2015)) was applied to deconvolute the transcriptomic dataset of graft biopsies of renal transplant patients without circulating DSA. In line with the hypothesis, patients with the higher number of NK cell in biopsy showed inferior graft survival (Figure 8A).

Histological analyses of graft biopsies from patients with missing-self induced NK cell- mediated rejection confirmed that mTORCl pathway was activated in NK cells adherent to graft microvasculature.

The ability of mTOR inhibitors to block mTORCl pathway and suppress missing-self induced cytotoxicity of human NK cells was then evaluated ex vivo. NK cells were purified from the circulation of 24 patients before and 1 month after introduction of everolimus and the level of phosphorylation of S6 Ribosomal Protein (S6RP, which is located downstream mTORCl) was measured by flow cytometry. Exposition to the drug in vivo, not only decreased the baseline level of phosphorylation of S6RP in NK cells, but also drastically reduced their response to IL-15 (Figure 8B). As expected from previous results, adjunction of mTOR inhibitors to co-cultures of K562 cells and purified NK cells, also significantly reduced mis sing- self induced cytotoxicity of the latters (Figure 8C). In an attempt to further document the potential of mTOR inhibitors to alleviate missing self- induced NK cell-mediated rejection in the clinical setting, the database of the heart transplantation department of George Pompidou University Hospitals (Paris, France) was searched. A 40-year-old heart transplant recipient with a clinical picture compatible with the diagnosis of missing self-induced NK-mediated rejection was identified. Pathological analyses of his heart graft biopsies revealed persistent microvascular inflammation in the absence of circulating DSA (Figure 8D) and genetic analyses confirmed the existence of a missing self. Interestingly, this patient first received a calcineurin inhibitor-based triple maintenance immunosuppression during the first 7 months post-transplantation, before an mTOR inhibitor was added to the immunosuppressive regimen. Re-analysis of heart graft biopsies by a trained pathologist (J.P. Duong), blinded for the delay since transplantation and the immunosuppressive regimen, revealed that introduction of mTOR inhibitor (everolimus) correlated with a significant reduction in the severity of microvascular inflammation (Figure 8D).

A 39-year-old renal transplant recipient with microvascular inflammation lesions on 3-month systematic graft biopsy in the absence of circulating DSA was prospectively identified in Lyon University Hospital (Figure 8E). On the basis of the genetic analyses, which confirmed the existence of 2 missing self, an mTOR inhibitor (everolimus) was introduced in replacement of the antimetabolite for maintenance immunosuppression. The graft biopsy performed 6 month later showed the total disappearance of microvascular inflammation (Figure 8E).

These 2 cases confirm the fact that mTOR inhibitors may be used to prevent the development of missing self-induced NK-mediated rejection in the clinical setting.

References

Throughout this application, various references describe the state of the art to which this invention pertains. The disclosures of these references are hereby incorporated by reference into the present disclosure.

1. Reindl-Schwaighofer R, Heinzel A, Signorini L, Thaunat O, Oberbauer R. Mechanisms underlying human genetic diversity: consequence for antigraft antibody responses. Transplant international : official journal of the European Society for Organ Transplantation 2017.

2. Calne RY, White DJ, Thiru S, et al. Cyclosporin A in patients receiving renal allografts from cadaver donors. Lancet 1978;2: 1323-7.

3. Lamb KE, Lodhi S, Meier-Kriesche HU. Long-term renal allograft survival in the United States: a critical reappraisal. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2011;11:450-62.

4. Terasaki PI, Cai J. Humoral theory of transplantation: further evidence. Current opinion in immunology 2005;17:541-5.

5. Russell PS, Chase CM, Winn HJ, Colvin RB. Coronary atherosclerosis in transplanted mouse hearts. II. Importance of humoral immunity. Journal of immunology 1994; 152:5135- 41.

6. Terasaki PI, Ozawa M. Predicting kidney graft failure by HLA antibodies: a prospective trial. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2004;4:438-43.

7. Pouliquen E, Koenig A, Chen CC, et al. Recent advances in renal transplantation: antibody- mediated rejection takes center stage. F1000prime reports 2015;7:51.

8. Sellares J, de Freitas DG, Mengel M, et al. Understanding the causes of kidney transplant failure: the dominant role of antibody-mediated rejection and nonadherence. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2012;12:388-99.

9. Wiebe C, Gibson IW, Blydt-Hansen TD, et al. Evolution and clinical pathologic correlations of de novo donor- specific HLA antibody post kidney transplant. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2012;12: 1157-67. 10. Loupy A, Toquet C, Rouvier P, et al. Late Failing Heart Allografts: Pathology of Cardiac Allograft Vasculopathy and Association With Antibody-Mediated Rejection.

American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2016;16: 111-20.

11. Roux A, Bendib Le Lan I, Holifanjaniaina S, et al. Antibody-Mediated Rejection in Lung Transplantation: Clinical Outcomes and Donor-Specific Antibody Characteristics. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2016;16: 1216-28.

12. Cantarovich D, De Amicis S, Akl A, et al. Posttransplant donor- specific anti-HLA antibodies negatively impact pancreas transplantation outcome. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2011;11:2737-46.

13. Thaunat O, Badet L, Dubois V, Kanitakis J, Petruzzo P, Morelon E. Immunopathology of rejection: do the rules of solid organ apply to vascularized composite allotransplantation? Current opinion in organ transplantation 2015;20:596- 601.

14. Chen CC, Pouliquen E, Broisat A, et al. Endothelial chimerism and vascular sequestration protect pancreatic islet grafts from antibody-mediated rejection. The Journal of clinical investigation 2018;128:219-32.

15. Loupy A, Lefaucheur C, Vernerey D, et al. Complement-binding anti-HLA antibodies and kidney-allograft survival. The New England journal of medicine 2013;369: 1215-26.

16. Sicard A, Ducreux S, Rabeyrin M, et al. Detection of C3d-binding donorspecific anti- HLA antibodies at diagnosis of humoral rejection predicts renal graft loss. Journal of the American Society of Nephrology : JASN 2015;26:457-67.

17. Guidicelli G, Guerville F, Lepreux S, et al. Non-Complement-Binding De Novo Donor- Specific Anti-HLA Antibodies and Kidney Allograft Survival. Journal of the American

Society of Nephrology : JASN 2016;27:615-25.

18. Hirohashi T, Uehara S, Chase CM, et al. Complement independent antibodymediated endarteritis and transplant arteriopathy in mice. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2010;10:510-7.

19. Fishbein GA, Fishbein MC. Morphologic and immunohistochemical findings in antibody- mediated rejection of the cardiac allograft. Human immunology 2012;73: 1213-7.

20. Gupta A, Broin PO, Bao Y, et al. Clinical and molecular significance of microvascular inflammation in transplant kidney biopsies. Kidney international 2016;89:217-25. 21. Haas M, Loupy A, Lefaucheur C, et al. The Banff 2017 Kidney Meeting Report: Revised diagnostic criteria for chronic active T cell-mediated rejection, antibodymediated rejection, and prospects for integrative endpoints for next- generation clinical trials. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2018;18:293-307.

22. Sicard A, Meas-Yedid V, Rabeyrin M, et al. Computer-assisted topological analysis of renal allograft inflammation adds to risk evaluation at diagnosis of humoral rejection. Kidney international 2017;92:214-26.

23. Le Bas-Bernardet S, Hourmant M, Coupel S, Bignon JD, Soulillou JP, Charreau B. Non- HLA-type endothelial cell reactive alloantibodies in pre-transplant sera of kidney recipients trigger apoptosis. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2003;3: 167-77.

24. Marcais A, Marotel M, Degouve S, et al. High mTOR activity is a hallmark of reactive natural killer cells and amplifies early signaling through activating receptors. eLife 2017;6. 25. Dragun D, Catar R, Philippe A. Non-HLA antibodies against endothelial targets bridging alio- and autoimmunity. Kidney international 2016;90:280-8.

26. Jackson AM, Sigdel TK, Delville M, et al. Endothelial cell antibodies associated with novel targets and increased rejection. Journal of the American Society of Nephrology : JASN 2015;26: 1161-71.

27. Sicard A, Chen CC, Morelon E, Thaunat O. Alloimmune- induced intragraft lymphoid neogenesis promotes B-cell tolerance breakdown that accelerates chronic rejection. Current opinion in organ transplantation 2016;21:368-74.

28. Thaunat O, Graff-Dubois S, Fabien N, et al. A stepwise breakdown of B-cell tolerance occurs within renal allografts during chronic rejection. Kidney international 2012;81:207-19. 29. Smith RN, Kawai T, Boskovic S, et al. Four stages and lack of stable accommodation in chronic alloantibody-mediated renal allograft rejection in Cynomolgus monkeys. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2008;8: 1662-72.

30. Hidalgo LG, Sis B, Sellares J, et al. NK cell transcripts and NK cells in kidney biopsies from patients with donor- specific antibodies: evidence for NK cell involvement in antibody- mediated rejection. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2010;10: 1812- 22. 31. Thielens A, Vivier E, Romagne F. NK cell MHC class I specific receptors (KIR): from biology to clinical intervention. Current opinion in immunology 2012;24:239-45.

32. Campbell KS, Purdy AK. Structure/function of human killer cell immunoglobulin-like receptors: lessons from polymorphisms, evolution, crystal structures and mutations. Immunology 2011;132:315-25.

33. Parham P. The genetic and evolutionary balances in human NK cell receptor diversity. Seminars in immunology 2008;20:311-6.

34. Shifrin N, Raulet DH, Ardolino M. NK cell self tolerance, responsiveness and missing self recognition. Seminars in immunology 2014;26: 138-44.

35. Kusunoki Y, Kyoizumi S, Honma M, et al. NK-mediated elimination of mutant lymphocytes that have lost expression of MHC class I molecules. Journal of immunology 2000;165:3555-63.

36. Wiertz EJ, Jones TR, Sun L, Bogyo M, Geuze HJ, Ploegh HL. The human cytomegalovirus US 11 gene product dislocates MHC class I heavy chains from the endoplasmic reticulum to the cytosol. Cell 1996;84:769-79.

37. Wagner JA, Rosario M, Romee R, et al. CD56bright NK cells exhibit potent antitumor responses following IL-15 priming. The Journal of clinical investigation 2017;127:4042-58.

38. Shaw J, Kollnberger S. New perspectives on the ligands and function of the killer cell immunoglobulin-like receptor KIR3DL2 in health and disease. Frontiers in immunology 2012;3:339.

39. Yawata M, Yawata N, Draghi M, Partheniou F, Little AM, Parham P. MHC class I- specific inhibitory receptors and their ligands structure diverse human NK-cell repertoires toward a balance of missing self-response. Blood 2008;112:2369-80.

40. Marcais A, Cherfils-Vicini J, Viant C, et al. The metabolic checkpoint kinase mTOR is essential for IL-15 signaling during the development and activation of NK cells. Nature immunology 2014;15:749-57.

41. Ekberg H, Tedesco-Silva H, Demirbas A, et al. Reduced exposure to calcineurin inhibitors in renal transplantation. The New England journal of medicine 2007;357:2562-75.

42. Thaunat O, Koenig A, Leibler C, Grimbert P. Effect of Immunosuppressive Drugs on Humoral AUosensitization after Kidney Transplant. Journal of the American Society of

Nephrology : JASN 2016;27: 1890-900.

43. Li J, Kim SG, Blenis J. Rapamycin: one drug, many effects. Cell metabolism 2014;19:373-9. 44. Joosten SA, van Dixhoorn MG, Borrias MC, et al. Antibody response against perlecan and collagen types IV and VI in chronic renal allograft rejection in the rat. The American journal of pathology 2002;160: 1301-10.

45. Kreijveld E, van der Meer A, Tijssen HJ, Hilbrands LB, Joosten I. KIR gene and KIR ligand analysis to predict graft rejection after renal transplantation. Transplantation

2007;84: 1045-51.

46. Tran TH, Mytilineos J, Scherer S, Laux G, Middleton D, Opelz G. Analysis of KIR ligand incompatibility in human renal transplantation. Transplantation 2005;80: 1121-3.

47. Laffont S, Seillet C, Ortaldo J, Coudert JD, Guery JC. Natural killer cells recruited into lymph nodes inhibit alloreactive T-cell activation through perforinmediated killing of donor allogeneic dendritic cells. Blood 2008;112:661-71.

48. Martinez-Llordella M, Lozano JJ, Puig-Pey I, et al. Using transcriptional profiling to develop a diagnostic test of operational tolerance in liver transplant recipients. The Journal of clinical investigation 2008;118:2845-57.

49. Yu G, Xu X, Vu MD, Kilpatrick ED, Li XC. NK cells promote transplant tolerance by killing donor antigen-presenting cells. The Journal of experimental medicine 2006;203: 1851- 8.

50. Grimbert P, Thaunat O. mTOR inhibitors and risk of chronic antibody-mediated rejection after kidney transplantation: where are we now? Transplant international : official journal of the European Society for Organ Transplantation 2017;30:647-57.

51. Kamar N, Del Bello A, Congy-Jolivet N, et al. Incidence of donor- specific antibodies in kidney transplant patients following conversion to an everolimus-based calcineurin inhibitor- free regimen. Clinical transplantation 2013;27:455-62.

52. Liefeldt L, Brakemeier S, Glander P, et al. Donor- specific HLA antibodies in a cohort comparing everolimus with cyclosporine after kidney transplantation. American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons 2012;12: 1192-8.

Claims

CLAIMS:

1. An in vitro method for predicting the risk of chronic transplant rejection in a subject who is the recipient of a transplant from a transplant donor, said method comprising:

the detection of missing-self activation of Natural Killer (NK) cells, wherein the detection of missing-self activation of NK cells is indicative of a risk for transplant rejection.

2. The method of claim 1, wherein the detection of missing-self activation of NK cells is achieved by the identification of the absence in the transplant donor of a ligand for a functional inhibitory Killer-cell Immunoglobulin-like Receptor (KIR) of the recipient subject.

3. The method according to claim 1, wherein the detection of missing-self activation of NK cells is achieved by the identification of the absence of expression by the endothelial cells of the transplant, of a ligand for a functional inhibitory KIR of the recipient subject.

4. The method according to claim 1, 2 or 3, wherein the detection of missing-self activation of NK cells comprises:

5. The method of claim 1, 2 or 3, wherein the detection of mis sing- self activation of NK cells comprises

- co-cultivating stimulated NK cells from the recipient subject with endothelial cells, preferably primary endothelial cells; and

- detecting the presence of a de granulation compound, preferably CD 107a.

6. The method of any one of claims 1-5, wherein the recipient subject is a candidate recipient subject.

7. A method of selecting a HLA compatible donor transplant for a candidate recipient subject comprising: - obtaining the inhibitory KIR genotype of the candidate recipient subject and the HLA-I genotype of the transplant donor and the candidate recipient subject;

- comparing the HLA-I genotype and the KIR genotype of the candidate recipient subject et determining the functional KIRs; and

8. The method of any one of claims 1-7, further comprising determining the proportion of NK cells in the recipient subject bearing inhibitory functional KIRs directed against HLA-I ligand missing in the donor HLA-I genotype.

9. The method of any one of claims 1-8, wherein the inhibitory KIRs comprise KIR2DL1, KIR2DL3, KIR2DL5, KIR3DL1, KIR3DL2 and KIR3DL3.

10. A Method for selecting a compatible donor transplant for a candidate recipient subject comprising:

- selecting a donor transplant which does not expresses the ligands expressed by the target cells for which a cell-death was induced after the co-cultivating step.

11. The method of any one of claims 1-9, wherein the transplant is a solid transplant organ, notably selected from heart transplant, lung transplant, kidney transplant, liver transplant, pancreas transplant, intestine transplant, thymus transplant and is preferably kidney transplant.

12- An mTorCl inhibitor, for use in the prevention or the treatment of chronic transplant rejection in a subject who is the recipient of a transplant and selected as having at least one recipient functional KIR directed against HLA-I ligand missing in donor HLA-I genotype.

13. An mTorCl inhibitor for use according to claim 12, wherein the inhibitor is rapamycin or one of its rapalog analog.