Role of αβ T Cell Depletion in Prevention of Graft versus Host Disease

Graft versus host disease (GVHD) represents a major complication of allogeneic hematopoietic stem cell transplantation (allo HCT). Graft cellular manipulation has been used to mitigate the risk of GVHD. The αβ T cells are considered the primary culprit for causing GVHD therefore depletion of this T cell subset emerged as a promising cellular manipulation strategy to overcome the human leukocyte antigen (HLA) barrier of haploidentical (haplo) HCT. This approach is also being investigated in HLA-matched HCT. In several studies, αβ T cell depletion HCT has been performed without pharmacologic GVHD prophylaxis, thus unleashing favorable effect of donor’s natural killer cells (NK) and γδ T cells. This article will discuss the evolution of this method in clinical practice and the clinical outcome as described in different clinical trials.


Introduction
Allogeneic hematopoietic stem cell transplantation (allo HCT) is a treatment modality for a variety of malignant and non-malignant diseases [1]. Only about 30% of patients have an HLA-matched sibling and in about 16-75% (depending on ethnicity), an HLA-matched unrelated donor can be identified [1]. In the absence of an HLA-matched sibling or volunteer donor, alternative donor of hematopoietic stem cells (HSC), such as unrelated umbilical cord blood (UCB) [2][3][4][5] or haplo donor [6][7][8][9][10][11] can be utilized. The main advantages of UCB products are the immediate availability and low risk of graft versus host disease (GVHD) with mismatched products (immature lymphocyte content). The disadvantages are high cost (30,000-40,000 US dollars per unit), difficult finding of units with adequate HSC for adults (in particular for overweight/obese recipients), delayed engraftment and immune reconstitution, and lack of the donor lymphocyte infusion (DLI) option. On the other hand, the main advantages of haplo HSC products are the availability of donors in most cases (first degree relative either parent, sibling or a child), relatively lower cost, faster engraftment (compared to UCB), and availability of DLI option.
The initial successful haplo HCT relied primarily on T cell depletion to control the high risk of GVHD induced by the high HLA disparity in this setting [12,13]. This approach also allowed elimination of post-transplant immunosuppression therapy (IST) which may unleash T cell-mediated graft versus tumor (GVT) effect. This approach would also avert several side effects related to the use of IST such as high risk of posterior reversible encephalopathy syndrome which is more pronounced in sickle cell patients receiving calcineurin inhibitors [14]. T cell depletion techniques have been refined

Rationale of αβ T Cell Depleted Hematopoietic Stem Cell Transplantation
Pre-clinical models of GVHD demonstrated that CD4+ and CD8+ T cells (= αβ T cells) to be major players in GVHD pathogenesis [37][38][39]. This causative correlation is the rationale for the use of αβ T cell depletion (rather than pan T cell depletion) allo HCT. The αβ T cell depletion is often combined with CD19+ B cell depletion for same reason explained above. The selective depletion of the αβ T cell from the infused graft spares γδ T cells and NK cells and likely favor their homeostatic reconstitution, thus potentially resulting in lower risk of infection [40,41] and relapse [42,43]. NK cells play a pivotal role in the defense against malignant transformed or virus-infected cells [44]. Allo-reactive NK cells have also been shown to positively affect the outcome of HCT via displaying GVT effect in children and adults without increasing risk of GVHD [8,[45][46][47][48][49][50][51]. In murine models, NK allo-reactive cells were able to kill host dendritic cells (one of the antigen presenting cells = APCs), and this can contribute to reducing the risk of GVHD, since recipient APCs are known to play a major role in GVHD pathophysiology [52]. The γδ T cells population is a component of the innate immune system. They can directly recognize self-expressed stress-related (e.g., viral or oncogenic) antigen on the cell surface triggering immediate cytotoxic effect [53][54][55]. This is in distinction to the limited capability of the αβ T cells and NK cells that can only recognize MHC-related peptides of tumor-associated antigens. Several preclinical and clinical observations have suggested the antineoplastic effect of γδ T cell against hematological malignancies [56,57] and solid tumors [58,59]. These data have been corroborated in clinical studies showing improved relapse-free survival with higher post-transplant γδ T cell counts in the peripheral blood [42,60,61]. For example, one study has shown that higher γδ T cell (≥ 10% of total lymphocytes) in the peripheral blood in earlier post-transplant time (between 2-9 months) was an independent factor for improved DFS [61]. The γδ T cells, alike NK cells, have not been implicated in causing GVHD [62][63][64]. Moreover, the γδ T cells were shown to facilitate engraftment of allogeneic stem cells in preclinical models [65,66]. This favorable effect on engraftment was also suggested by clinical observation [67,68]. It is to be noted that despite the hypothesized favorable outcome of using αβ T cell depletion transplant, this approach was not directly compared to the traditional pan T cell depletion. Only Lang et al. [69] reported improved T and NK cell recovery following αβ T cell depletion transplant when compared to historical cases of pan T cell depletion.

Technical Methods
The HSC product contains a variety of cells including myeloid precursors and lymphocytes in addition to the minor component (~1%) of stem cells ( Figure 2) [70]. Various methods have been employed for ex vivo T cell depletion and reviewed in recent literature [71]. The earliest clinical method involved the use of soybean lectin agglutination with T cell resetting with sheep red blood cells [72]. Subsequently, T cell monoclonal antibodies (in combination with immunotoxins or complement) were used [73]. The addition of complement or immunotoxins to the anti-T cell

Rationale of αβ T Cell Depleted Hematopoietic Stem Cell Transplantation
Pre-clinical models of GVHD demonstrated that CD4+ and CD8+ T cells (=αβ T cells) to be major players in GVHD pathogenesis [37][38][39]. This causative correlation is the rationale for the use of αβ T cell depletion (rather than pan T cell depletion) allo HCT. The αβ T cell depletion is often combined with CD19+ B cell depletion for same reason explained above. The selective depletion of the αβ T cell from the infused graft spares γδ T cells and NK cells and likely favor their homeostatic reconstitution, thus potentially resulting in lower risk of infection [40,41] and relapse [42,43]. NK cells play a pivotal role in the defense against malignant transformed or virus-infected cells [44]. Allo-reactive NK cells have also been shown to positively affect the outcome of HCT via displaying GVT effect in children and adults without increasing risk of GVHD [8,[45][46][47][48][49][50][51]. In murine models, NK allo-reactive cells were able to kill host dendritic cells (one of the antigen presenting cells = APCs), and this can contribute to reducing the risk of GVHD, since recipient APCs are known to play a major role in GVHD pathophysiology [52]. The γδ T cells population is a component of the innate immune system. They can directly recognize self-expressed stress-related (e.g., viral or oncogenic) antigen on the cell surface triggering immediate cytotoxic effect [53][54][55]. This is in distinction to the limited capability of the αβ T cells and NK cells that can only recognize MHC-related peptides of tumor-associated antigens. Several preclinical and clinical observations have suggested the antineoplastic effect of γδ T cell against hematological malignancies [56,57] and solid tumors [58,59]. These data have been corroborated in clinical studies showing improved relapse-free survival with higher post-transplant γδ T cell counts in the peripheral blood [42,60,61]. For example, one study has shown that higher γδ T cell (≥10% of total lymphocytes) in the peripheral blood in earlier post-transplant time (between 2-9 months) was an independent factor for improved DFS [61]. The γδ T cells, alike NK cells, have not been implicated in causing GVHD [62][63][64]. Moreover, the γδ T cells were shown to facilitate engraftment of allogeneic stem cells in preclinical models [65,66]. This favorable effect on engraftment was also suggested by clinical observation [67,68]. It is to be noted that despite the hypothesized favorable outcome of using αβ T cell depletion transplant, this approach was not directly compared to the traditional pan T cell depletion. Only Lang et al. [69] reported improved T and NK cell recovery following αβ T cell depletion transplant when compared to historical cases of pan T cell depletion.

Technical Methods
The HSC product contains a variety of cells including myeloid precursors and lymphocytes in addition to the minor component (~1%) of stem cells ( Figure 2) [70]. Various methods have been employed for ex vivo T cell depletion and reviewed in recent literature [71]. The earliest clinical method involved the use of soybean lectin agglutination with T cell resetting with sheep red blood cells [72]. Subsequently, T cell monoclonal antibodies (in combination with immunotoxins or complement) were used [73]. The addition of complement or immunotoxins to the anti-T cell antibody is essential for elimination of the T cells. This was recognized after encountering high risk of GVHD with the earlier use of T cell monoclonal antibody alone [74]. The discovery of T10B9 by University of Kentucky (USA) allowed the selective depletion of the αβ TC [75]. The procedure of αβ T cell depletion has been described before [76][77][78]. An updated report of the procedure efficiency has also been published [76].
antibody is essential for elimination of the T cells. This was recognized after encountering high risk of GVHD with the earlier use of T cell monoclonal antibody alone [74]. The discovery of T10B9 by University of Kentucky (USA) allowed the selective depletion of the αβ TC [75]. The procedure of αβ T cell depletion has been described before [76][77][78]. An updated report of the procedure efficiency has also been published [76]. In summary, the graft processing for αβ T cell depletion is done using the CliniMACS device ® TCRαβ-Biotin system (Miltenyi Biotec, Bergisch Gladbach, Germany). The allogeneic donors are mobilized with filgrastim G-CSF for 4 days with leukapheresis starting on day 5 (and possibly day 6) per standard guidelines [79]. Peripheral blood CD34+ cell count is checked on the day of apheresis (day 5). A count of ≥ 40/µL is predictive of an adequate collection in one apheresis session, while a count < 20/µL often predict suboptimal collection (even in 2 sessions). In these donors, plerixafor is considered as described previously [80]. However, it is to be noted that plerixafor is not currently approved for this indication (volunteer donor) by the Food and Drug Administration in the USA. The target number of CD34+ stem cells in the apheresis product (i.e., prior to αβ T cell depletion) for pediatric population is 40 (minimum of 12-15) × 10 6 cells/kg recipient weight. The leukapheresis product then undergoes negative selection (i.e., depletion) of the αβ T cells prior to infusion to the patient. This depletion typically results in ~20 (minimum of 8-10) × 10 6 cells/kg CD34+ cells (i.e., allowing for up to 40% loss during the depletion procedure). Prior to the immunomagnetic labeling of the apheresis product, it is washed to remove platelets and the cell concentration will be adjusted in preparation for antibody labeling. The apheresis product is then labeled using the CliniMACS TCRαβ Biotin kit (Miltenyi Biotec, Bergisch Gladbach, Germany) and CD19+ immunomagnetic microbeads. After immunomagnetic labeling, the cells are washed to remove unbound microbeads ( Figure 3). The labeled product is loaded onto the CliniMACS device where labeled cells are depleted and the negative fraction is eluted off the device. This negative fraction is then centrifuged and volume-reconstituted to obtain the final product. We do not have a maximum limit of CD34+ cells to be infused, however, we target a maximum dose of αβ T cells of 1 × 10 5 /kg at the end of the negative depletion procedure. If the residual number of αβ T cells is > 1 × 10 5 /kg, a selected part of the product can be eliminated and cryopreserved. If this exclusion compromises the minimum CD34+ stem cells, we perform CD34+ cell selection on that part of the graft. Our transplant protocol typically involves using rituximab at day +1 to eradicate residual B cell in the product unless the CD19+ B cells in the final product is < 1 × 10 5 CD19+ cells/kg. In summary, the graft processing for αβ T cell depletion is done using the CliniMACS device ® TCRαβ-Biotin system (Miltenyi Biotec, Bergisch Gladbach, Germany). The allogeneic donors are mobilized with filgrastim G-CSF for 4 days with leukapheresis starting on day 5 (and possibly day 6) per standard guidelines [79]. Peripheral blood CD34+ cell count is checked on the day of apheresis (day 5). A count of ≥ 40/µL is predictive of an adequate collection in one apheresis session, while a count < 20/µL often predict suboptimal collection (even in 2 sessions). In these donors, plerixafor is considered as described previously [80]. However, it is to be noted that plerixafor is not currently approved for this indication (volunteer donor) by the Food and Drug Administration in the USA. The target number of CD34+ stem cells in the apheresis product (i.e., prior to αβ T cell depletion) for pediatric population is 40 (minimum of 12-15) × 10 6 cells/kg recipient weight. The leukapheresis product then undergoes negative selection (i.e., depletion) of the αβ T cells prior to infusion to the patient. This depletion typically results in~20 (minimum of 8-10) × 10 6 cells/kg CD34+ cells (i.e., allowing for up to 40% loss during the depletion procedure). Prior to the immunomagnetic labeling of the apheresis product, it is washed to remove platelets and the cell concentration will be adjusted in preparation for antibody labeling. The apheresis product is then labeled using the CliniMACS TCRαβ Biotin kit (Miltenyi Biotec, Bergisch Gladbach, Germany) and CD19+ immunomagnetic microbeads. After immunomagnetic labeling, the cells are washed to remove unbound microbeads (Figure 3). The labeled product is loaded onto the CliniMACS device where labeled cells are depleted and the negative fraction is eluted off the device. This negative fraction is then centrifuged and volumereconstituted to obtain the final product. We do not have a maximum limit of CD34+ cells to be infused, however, we target a maximum dose of αβ T cells of 1 × 10 5 /kg at the end of the negative depletion procedure. If the residual number of αβ T cells is >1 × 10 5 /kg, a selected part of the product can be eliminated and cryopreserved. If this exclusion compromises the minimum CD34+ stem cells, we perform CD34+ cell selection on that part of the graft. Our transplant protocol typically involves using rituximab at day +1 to eradicate residual B cell in the product unless the CD19+ B cells in the final product is <1 × 10 5 CD19+ cells/kg. We sometimes use BM product if the donor is a child or an adult donor declines PBSC apheresis. However, it is to be noted that, in order to optimize the selection process, the maximum volume of packed red blood cells (RBCs) in the product (pre-selection) that is allowed to go on the CliniMACS column is 30 mL (i.e., 100 mL of BM with 30% HCT). Therefore, BM product is RBC-depleted using Ficoll ® density gradient separation (GE Healthcare Bio-Sciences, Pittsburgh, PA, USA) prior to proceeding with selection on the CliniMACS device. The RBC-reduced product is stored at +1 to +8 °C until used.

Clinical Outcome of ΑΒ T Cells Depletion HCT
The utilization of αβ T cells depletion in allogenic HCT has been evaluated in treating both malignant and non-malignant etiologies [81]. The majority of the published studies have been conducted in pediatric haplo HCT setting [61,82]. Selected seven studies are summarized and discussed below (Table 1) [69,[83][84][85][86][87][88]. Transplant outcome using this approach has also been described in other studies [28,29,82]. Several other individual case reports have reported the use of αβ T cells depletion in different non-malignant conditions as Wiskott-Aldrich, β thalassemia and Hoyeraal-Hreidarsson syndrome with favorable outcome [89][90][91]. We sometimes use BM product if the donor is a child or an adult donor declines PBSC apheresis. However, it is to be noted that, in order to optimize the selection process, the maximum volume of packed red blood cells (RBCs) in the product (pre-selection) that is allowed to go on the CliniMACS column is 30 mL (i.e., 100 mL of BM with 30% HCT). Therefore, BM product is RBC-depleted using Ficoll ® density gradient separation (GE Healthcare Bio-Sciences, Pittsburgh, PA, USA) prior to proceeding with selection on the CliniMACS device. The RBC-reduced product is stored at +1 to +8 • C until used.

Clinical Outcome of AB T Cells Depletion HCT
The utilization of αβ T cells depletion in allogenic HCT has been evaluated in treating both malignant and non-malignant etiologies [81]. The majority of the published studies have been conducted in pediatric haplo HCT setting [61,82]. Selected seven studies are summarized and discussed below (Table 1) [69,[83][84][85][86][87][88]. Transplant outcome using this approach has also been described in other studies [28,29,82]. Several other individual case reports have reported the use of αβ T cells depletion in different non-malignant conditions as Wiskott-Aldrich, β thalassemia and Hoyeraal-Hreidarsson syndrome with favorable outcome [89][90][91].

Engraftment and Immune Reconstitution
The engraftment failure rate was variable among different studies ranging from 0% reported by Maschan et al. [85], and up to 27% by Balashov et al. [84]. The median day for neutrophil and platelet count recovery reported in these studies ranged from 12-16 and 10-14 days respectively. The recovery of γδ T cells preceded αβ T cells with a median of 7-10 days. Two studies reported similar results of immune reconstitution with T cell > 500/µL and B cell > 200/µL on day +120 [69,84]. Another study reported similar data for T cell recovery at day 120 but delayed B cell recovery >150/µL until 6 months [83].

Graft Versus Host Disease (GVHD)
Different studies used different prophylaxis regimens against GVHD. Antithymocyte globulin (ATG) was used in all studies as part of the preparative regimen. Only one study used OKT3 instead of ATG before 2012 in a subset of 34 patients [69]. While a German group [69] used ATG distal to day 0 (day −9 to −12) with the primary purpose of prevention of graft failure, an Italian group [83] used it more proximal to day 0 (day −3 to −5) in order to prevent both graft failure and GVHD hypothesizing that this will not influence the post-transplant recovery of γδ T cells which is expected to occur after biological elimination of ATG. No pharmacologic GVHD prophylaxis was used by an Italian study [83]. Three studies used single agent mycophenolate mofetil [69,86,87] and another study used tacrolimus in all patients plus methotrexate (n = 34), mycophenolate mofetil (n = 2), or cyclosporine (n = 1) [84]. In another study, single agent tacrolimus (n = 2), methotrexate (n = 5) or both (n = 21) were used while 5 patients did not receive pharmacologic prophylaxis [85]. B cell depletion was done simultaneously in most studies via ex vivo CD 19 depletion (along with αβ T cell depletion). Two studies used rituximab for in vivo B-cell depletion [83,84].
The rate of acute and chronic GVHD (aGVHD and cGVHD) was variable among studies, but occurred at low rate and was mostly low grade. The lowest incidence of aGVHD of 3% reported by one study [86]. Bertaina et al. reported 13% risk of aGVHD (mostly grade I-II and only skin involvement) with no reported cGVHD with a median follow-up of 18 months [83]. In another study, the risk of aGVHD was 25% (with 15% risk of grade III) with a cGVHD risk of 27% (extensive disease of 9%) [69]. In the adult cohort by Kaynar et al. 38% developed aGVHD (grade I-II was 27%), and 6% developed cGVHD (2 patients; one was extensive) [87]. In the study by Balashov et al. 23% developed aGVHD (with one patient with grade IV that turned into refractory cGVHD) [84]. The highest incidence of grade II-III aGVHD was reported by Maschan et al. as 39% (none developed grade IV) with risk of cGVHD of 30% with a median follow up of 2 years (some patients received donor lymphocyte infusion) [85]. In the report by Gonzalez et al. risk of aGVHD and cGVHD was 18% and 14% respectively [88].

Relapse and Survival
The relapse of malignancy post-transplant was the major cause of mortality. Relapse rates are ranging from 22-58% while relapse-related mortality rates 19

Conclusions and Future Perspective
These clinical data are suggestive of a promising role of αβ T cell depletion to overcome the HLA disparity haplo HCT. Although this approach is adopted by several European centers, it has not gotten a wide utilization in the USA except for few pediatric centers. Several ongoing studies are under way using either haplo or HLA-matched HCT ( Table 2). Comparative studies are lacking to compare αβ T cell depletion HCT and other modalities of haplo HCT such as Pan T cell depletion or PTCy. Clinical trials evaluating the therapeutic utility of γδ T cells for hematological malignancies are lacking. An ongoing phase I trial is underway to evaluate the safety and feasibility of infusing add-back αβ T cell-depleted product after haplo HCT (NCT02193880). Suicide gene (caspase-9) programming of the add-back T cells has been used in order to eliminate the T cells (via therapeutic activation of the suicide gene) in case severe GVHD develops [92]. This approach is currently under investigation (NCT01744223). Acknowledgments: Disclosure: Ayman Saad discloses grant support (American Porphyria foundation), consultation (Medpace Inc), research support (Astellas and Fate Therapeutics), honoraria (Alxion, and Spectrum), and royalty for licensing of intellectual property (Incysus Biomedical).
Author Contributions: Haitham Abdelhakim collected data on outcome, Hisham Abdel-Azim collected data of technical methods and literature review, Ayman Saad did comprehensive literature review.