- Email: [email protected]
Gene Editing Expands the Donor Pool for CCR5-Negative Stem Cell Transplants
Cell Stem Cell
In Translation Gene Editing Expands the Donor Pool for CCR5-Negative Stem Cell Transplants Paula M. Cannon1,* 1Keck School of Medicine of the University of Southern California, Los Angeles, CA, USA *Correspondence: [email protected] https://doi.org/10.1016/j.stem.2019.11.006
Cell therapy efforts for treating HIV+ patients are challenged by limited availability of donors with naturally occurring CCR5 mutations conferring resistance. Xu et al. (2019) report a CRISPR-based method for disrupting CCR5 in hematopoietic stem cells prior to transplant, providing a proof of concept for expanding the pool of potential donors. The news last year that CRISPR-Cas9 gene editing had been used to create genetically modified twin baby girls (Cyranoski and Ledford, 2018) catapulted discussion of the ethics of gene editing into the public domain. It also made familiar the concept that editing the CCR5 gene, a major entry co-receptor used by the human immunodeficiency virus (HIV), is being considered as an approach to blocking HIV. Xu et al. (2019) now report early clinical results from a more acceptable application of CRISPR gene editing: disrupting CCR5 in donor hematopoietic stem and progenitor cells (HSPCs) prior to transplantation into an HIV+ adult. The rationale for editing CCR5 in HSPCs derives from the cases of the Berlin and London Patients, who were cured of HIV after undergoing allogeneic HSPC €tter transplantations (allo-HSCTs) (Hu et al., 2009; Gupta et al., 2019). Like the patient in Xu et al., these individuals underwent allo-HSCTs as treatment for hematologic malignancies. However, the tissue-matched donors for the London and Berlin Patients were also homozygous for the defective CCR5D32 mutation. In this way, a bonus of their cancer treatment was that they also received the HIV-resistant immune system of their donors. The striking successes of these two CCR5D32 allo-HSCTs suggests that HIV+ blood cancer patients requiring an allo-HSCT should also consider using a CCR5D32 donor, and efforts are underway to build databases of potential donors. Unfortunately, the relatively low frequency (1%) of CCR5D32 homozygotes means that finding donors who are both tissue-matched and CCR5D32 homozygous is a significant challenge. In an elegant twist, Xu et al. replaced the need
to identify a CCR5D32 homozygote by gene editing HSPCs from their patient’s tissue-matched donor. Allo-HSCTs have several features that may be important for an effective antiHIV therapy. The conditioning regimen used, combined with a graft versus host (GvH) effect from the donor-derived immune system, contributes to removal of residual cancer cells. These same effects can also deplete the reservoir of HIV-infected cells that persist in patients, even when they are on antiretroviral therapy (ART). However, the relative contribution of these three major elements of the CCR5D32 cures—conditioning, GvH effects, and transplantation with HIV-resistant cells—is currently unclear. What is known is that conditioning and GvH alone are not sufficient for a cure, since although allo-HSCTs using CCR5 wildtype donors can significantly shrink the HIV reservoir, allowing extended periods of time off ART, the virus eventually rebounds (Henrich et al., 2014). The approach of Xu et al. represents a next step in this evolving strategy by testing whether allo-HSCTs containing some portion of CCR5-edited cells can provide greater benefits than a CCR5 wild-type transplant (Figure 1). To test this idea, Xu et al. edited the donor CD34+ HSPCs before transplantation using CRISPR-Cas9 but achieved only modest levels of CCR5 disruption. The cell product had CCR5 allelic disruption rates of 17.8%, while rates in patient bone marrow cells and peripheral CD4+ T cells post-transplantation were 5.2%–8.28% and below 4.5%, respectively. Since only homozygous CCR5-disrupted cells would be HIV resistant, the frequency of resistant cells in their patient
is expected to be even lower. However, it was still possible to ask whether even these modest rates of CCR5 editing could impact HIV infection as the patient agreed to stop taking ART for a limited time period. ART interruptions typically cause viral rebound within 1 month and are currently the gold standard way to assess any impact on HIV disease. Following ART cessation, three scenarios could be imagined. First, no virus could rebound, providing a strong indication that the treatment was of sufficient potency that any recrudescence of the virus would be quickly limited by the lack of available CCR5+ immune cells. A second scenario could see virus rebound, but with delayed kinetics or reduced peak viremia, reflecting a partial protection. An appealing aspect of this second scenario is that the rebounding virus could itself select for CCR5-negative cells and potentially amplify the therapy. Finally, a third scenario would be no obvious impact on HIV replication, as was observed by Xu et al. While this lack of effect was perhaps not surprising giving the low levels of edited cells, it was still possible to ask if any hints of efficacy were present. In that regard the authors noted that the percentage of CCR5-edited alleles increased from 2.96% to 4.39% at a single time point during ART withdrawal. Although this may indeed have reflected a survival advantage of edited cells, the levels dropped back down by the next sample point within 10 days of restarting ART, when HIV levels were still quite high. This suggests that the observation could instead reflect sampling variations. In complementary approaches, other groups are evaluating whether providing
Cell Stem Cell 25, December 5, 2019 ª 2019 Elsevier Inc. 735
Cell Stem Cell
In Translation CCR5-edited autologous cells to lifetime ART. Addressing 1. Allogeneic HSCT alone could work (Figure 1). whether gene editing could Clinical trials are underway to re-create the success of the 2. Autologous test zinc-finger nuclease ediBerlin and London Patient T cells ted autologous T cells and cures, or whether this will (a) WT donor HSCT (Wang and Cannon, also work for more broadly 2016), and tantalizing hints of applicable autologous apcontrol have occurred in the proaches, awaits future trials early T cell trials (Tebas et al., with more efficiently edited 2014). Such autologous apcellular products. (b) CCR5 proaches provide the potential CCR5 ZFNs ACKNOWLEDGMENTS '32/'32 to expand the patient pool donor beyond the very limited group P.M.C. is supported by grant U19 of cancer patients undergoing HL129902 from the National Heart, allo-HSCT, since finding a Lung, and Blood Institute. donor is not an issue. HowevREFERENCES er, it is not yet known if the gentler conditioning regimens Cyranoski, D., and Ledford, H. appropriate for auto-HSCT in (c) WT (2018). Genome-edited baby claim non-cancer patients will prodonor 3. Autologous provokes international outcry. Nature 563, 607–608. vide the same benefits as HSPC allo-HSCT. Or indeed, if a parGupta, R.K., Abdul-Jawad, S., McCoy, L.E., Mok, H.P., Peppa, D., tial replacement with CCR5Figure 1. Different Approaches to Treating HIV using Cell Salgado, M., Martinez-Picado, J., edited cells will be as potent Transplants Nijhuis, M., Wensing, A.M.J., Lee, as the effects of a total Allo-HSCT has been performed using HSPCs from a normal (CCR5 WT) H., et al. (2019). HIV-1 remission donor, a CCR5D32 homozygous donor, or, in the approach described by Xu following CCR5D32/D32 haematoCCRD32 allo-HSCT. A future et al. (2019), using CCR5 CRISPR-edited cells from a WT donor. Autologous poietic stem-cell transplantation. comparison of outcomes HSPC and T cell trials are also underway using CCR5 ZFN-edited cells. To Nature 568, 244–248. from CCR5-edited auto and date, only the CCR5D32 allo-HSCT approach has been reported to cure HIV. Henrich, T.J., Hanhauser, E., Marty, allo-HSCT approaches should F.M., Sirignano, M.N., Keating, S., provide important insights into Lee, T.H., Robles, Y.P., Davis, B.T., the role of the various elements of the Ber- donor cells, although co-introducing or Li, J.Z., Heisey, A., et al. (2014). Antiretroviral-free remission and viral rebound after allogeneic lin and London cures. reserving non-edited HSPCs provides a HIV-1 stem cell transplantation: report of 2 cases. Ann. Beyond HIV applications, the work in Xu reasonable safety backup for cases Intern. Med. 161, 319–327. et al. provides important information where the failure of a transplant would €tter, G., Nowak, D., Mossner, M., Ganepola, S., Hu about the safety of CRISPR-edited be of grave concern. €ssig, A., Allers, K., Schneider, T., Hofmann, J., Mu CCR5 was the first clinical gene editing Ku€cherer, C., Blau, O., et al. (2009). Long-term HSPCs. They found that the edited allogeof HIV by CCR5 Delta32/Delta32 stemneic HSPCs could persist for at least target (Wang and Cannon, 2016) and control cell transplantation. N. Engl. J. Med. 360, 692–698. 19 months and produced mature progeny despite the recent CRISPR babies controwith the characteristic signatures of CCR5 versy, it remains a popular approach, not Tebas, P., Stein, D., Tang, W.W., Frank, I., Wang, S.Q., Lee, G., Spratt, S.K., Surosky, R.T., Giedlin, editing. In addition, comprehensive least because gene disruption is the M.A., Nichol, G., et al. (2014). Gene editing of genomic analyses showed no evidence simplest type of gene editing to achieve. CCR5 in autologous CD4 T cells of persons inof adverse or off-target editing events, As noted by Xu et al., its suitability was fected with HIV. N. Engl. J. Med. 370, 901–910. which is a major concern for gene editing recently questioned by a report that Wang, C.X., and Cannon, P.M. (2016). The clinical therapies. That said, the broader applica- CCR5D32 homozygotes have shorter life- applications of genome editing in HIV. Blood 127, 2546–2552. bility of their data does have limitations. spans (Wei and Nielsen, 2019), although The editing protocol used is unlikely to this paper has now been retracted. Xu Wei, X., and Nielsen, R. (2019). Retraction Note: CCR5-D32 is deleterious in the homozygous state be acceptable in the US since CRISPR- et al.’s finding of successful engraftment in humans. Nat. Med., in press. Published online Cas9 reagents were delivered using multi- and persistence of CRISPR-edited October 8, 2019. https://doi.org/10.1038/s41591ple small-scale reactions and research- HSPCs provides a welcome reassurance 019-0637-6. grade equipment. Technical problems of the safety and feasibility of somatic Xu, L., Wang, J., Liu, Y., Xie, L., Su, B., Mou, D., with isolation of the CD34+ cell product cell gene editing and represents impor- Wang, L., Liu, T., Wang, X., and Zhang, B. (2019). CRISPR-Edited Stem Cells in a Patient with HIV also meant that the final edited cell prod- tant progress on the path to developing and Acute Lymphocytic Leukemia. N. Engl. J. uct was diluted with non-edited bulk effective alternatives for HIV+ individuals Med. 381, 1240–1247.
736 Cell Stem Cell 25, December 5, 2019
In Translation Gene Editing Expands the Donor Pool for CCR5-Negative Stem Cell Transplants Paula M. Cannon1,* 1Keck School of Medicine of the University of Southern California, Los Angeles, CA, USA *Correspondence: [email protected] https://doi.org/10.1016/j.stem.2019.11.006
Cell therapy efforts for treating HIV+ patients are challenged by limited availability of donors with naturally occurring CCR5 mutations conferring resistance. Xu et al. (2019) report a CRISPR-based method for disrupting CCR5 in hematopoietic stem cells prior to transplant, providing a proof of concept for expanding the pool of potential donors. The news last year that CRISPR-Cas9 gene editing had been used to create genetically modified twin baby girls (Cyranoski and Ledford, 2018) catapulted discussion of the ethics of gene editing into the public domain. It also made familiar the concept that editing the CCR5 gene, a major entry co-receptor used by the human immunodeficiency virus (HIV), is being considered as an approach to blocking HIV. Xu et al. (2019) now report early clinical results from a more acceptable application of CRISPR gene editing: disrupting CCR5 in donor hematopoietic stem and progenitor cells (HSPCs) prior to transplantation into an HIV+ adult. The rationale for editing CCR5 in HSPCs derives from the cases of the Berlin and London Patients, who were cured of HIV after undergoing allogeneic HSPC €tter transplantations (allo-HSCTs) (Hu et al., 2009; Gupta et al., 2019). Like the patient in Xu et al., these individuals underwent allo-HSCTs as treatment for hematologic malignancies. However, the tissue-matched donors for the London and Berlin Patients were also homozygous for the defective CCR5D32 mutation. In this way, a bonus of their cancer treatment was that they also received the HIV-resistant immune system of their donors. The striking successes of these two CCR5D32 allo-HSCTs suggests that HIV+ blood cancer patients requiring an allo-HSCT should also consider using a CCR5D32 donor, and efforts are underway to build databases of potential donors. Unfortunately, the relatively low frequency (1%) of CCR5D32 homozygotes means that finding donors who are both tissue-matched and CCR5D32 homozygous is a significant challenge. In an elegant twist, Xu et al. replaced the need
to identify a CCR5D32 homozygote by gene editing HSPCs from their patient’s tissue-matched donor. Allo-HSCTs have several features that may be important for an effective antiHIV therapy. The conditioning regimen used, combined with a graft versus host (GvH) effect from the donor-derived immune system, contributes to removal of residual cancer cells. These same effects can also deplete the reservoir of HIV-infected cells that persist in patients, even when they are on antiretroviral therapy (ART). However, the relative contribution of these three major elements of the CCR5D32 cures—conditioning, GvH effects, and transplantation with HIV-resistant cells—is currently unclear. What is known is that conditioning and GvH alone are not sufficient for a cure, since although allo-HSCTs using CCR5 wildtype donors can significantly shrink the HIV reservoir, allowing extended periods of time off ART, the virus eventually rebounds (Henrich et al., 2014). The approach of Xu et al. represents a next step in this evolving strategy by testing whether allo-HSCTs containing some portion of CCR5-edited cells can provide greater benefits than a CCR5 wild-type transplant (Figure 1). To test this idea, Xu et al. edited the donor CD34+ HSPCs before transplantation using CRISPR-Cas9 but achieved only modest levels of CCR5 disruption. The cell product had CCR5 allelic disruption rates of 17.8%, while rates in patient bone marrow cells and peripheral CD4+ T cells post-transplantation were 5.2%–8.28% and below 4.5%, respectively. Since only homozygous CCR5-disrupted cells would be HIV resistant, the frequency of resistant cells in their patient
is expected to be even lower. However, it was still possible to ask whether even these modest rates of CCR5 editing could impact HIV infection as the patient agreed to stop taking ART for a limited time period. ART interruptions typically cause viral rebound within 1 month and are currently the gold standard way to assess any impact on HIV disease. Following ART cessation, three scenarios could be imagined. First, no virus could rebound, providing a strong indication that the treatment was of sufficient potency that any recrudescence of the virus would be quickly limited by the lack of available CCR5+ immune cells. A second scenario could see virus rebound, but with delayed kinetics or reduced peak viremia, reflecting a partial protection. An appealing aspect of this second scenario is that the rebounding virus could itself select for CCR5-negative cells and potentially amplify the therapy. Finally, a third scenario would be no obvious impact on HIV replication, as was observed by Xu et al. While this lack of effect was perhaps not surprising giving the low levels of edited cells, it was still possible to ask if any hints of efficacy were present. In that regard the authors noted that the percentage of CCR5-edited alleles increased from 2.96% to 4.39% at a single time point during ART withdrawal. Although this may indeed have reflected a survival advantage of edited cells, the levels dropped back down by the next sample point within 10 days of restarting ART, when HIV levels were still quite high. This suggests that the observation could instead reflect sampling variations. In complementary approaches, other groups are evaluating whether providing
Cell Stem Cell 25, December 5, 2019 ª 2019 Elsevier Inc. 735
Cell Stem Cell
In Translation CCR5-edited autologous cells to lifetime ART. Addressing 1. Allogeneic HSCT alone could work (Figure 1). whether gene editing could Clinical trials are underway to re-create the success of the 2. Autologous test zinc-finger nuclease ediBerlin and London Patient T cells ted autologous T cells and cures, or whether this will (a) WT donor HSCT (Wang and Cannon, also work for more broadly 2016), and tantalizing hints of applicable autologous apcontrol have occurred in the proaches, awaits future trials early T cell trials (Tebas et al., with more efficiently edited 2014). Such autologous apcellular products. (b) CCR5 proaches provide the potential CCR5 ZFNs ACKNOWLEDGMENTS '32/'32 to expand the patient pool donor beyond the very limited group P.M.C. is supported by grant U19 of cancer patients undergoing HL129902 from the National Heart, allo-HSCT, since finding a Lung, and Blood Institute. donor is not an issue. HowevREFERENCES er, it is not yet known if the gentler conditioning regimens Cyranoski, D., and Ledford, H. appropriate for auto-HSCT in (c) WT (2018). Genome-edited baby claim non-cancer patients will prodonor 3. Autologous provokes international outcry. Nature 563, 607–608. vide the same benefits as HSPC allo-HSCT. Or indeed, if a parGupta, R.K., Abdul-Jawad, S., McCoy, L.E., Mok, H.P., Peppa, D., tial replacement with CCR5Figure 1. Different Approaches to Treating HIV using Cell Salgado, M., Martinez-Picado, J., edited cells will be as potent Transplants Nijhuis, M., Wensing, A.M.J., Lee, as the effects of a total Allo-HSCT has been performed using HSPCs from a normal (CCR5 WT) H., et al. (2019). HIV-1 remission donor, a CCR5D32 homozygous donor, or, in the approach described by Xu following CCR5D32/D32 haematoCCRD32 allo-HSCT. A future et al. (2019), using CCR5 CRISPR-edited cells from a WT donor. Autologous poietic stem-cell transplantation. comparison of outcomes HSPC and T cell trials are also underway using CCR5 ZFN-edited cells. To Nature 568, 244–248. from CCR5-edited auto and date, only the CCR5D32 allo-HSCT approach has been reported to cure HIV. Henrich, T.J., Hanhauser, E., Marty, allo-HSCT approaches should F.M., Sirignano, M.N., Keating, S., provide important insights into Lee, T.H., Robles, Y.P., Davis, B.T., the role of the various elements of the Ber- donor cells, although co-introducing or Li, J.Z., Heisey, A., et al. (2014). Antiretroviral-free remission and viral rebound after allogeneic lin and London cures. reserving non-edited HSPCs provides a HIV-1 stem cell transplantation: report of 2 cases. Ann. Beyond HIV applications, the work in Xu reasonable safety backup for cases Intern. Med. 161, 319–327. et al. provides important information where the failure of a transplant would €tter, G., Nowak, D., Mossner, M., Ganepola, S., Hu about the safety of CRISPR-edited be of grave concern. €ssig, A., Allers, K., Schneider, T., Hofmann, J., Mu CCR5 was the first clinical gene editing Ku€cherer, C., Blau, O., et al. (2009). Long-term HSPCs. They found that the edited allogeof HIV by CCR5 Delta32/Delta32 stemneic HSPCs could persist for at least target (Wang and Cannon, 2016) and control cell transplantation. N. Engl. J. Med. 360, 692–698. 19 months and produced mature progeny despite the recent CRISPR babies controwith the characteristic signatures of CCR5 versy, it remains a popular approach, not Tebas, P., Stein, D., Tang, W.W., Frank, I., Wang, S.Q., Lee, G., Spratt, S.K., Surosky, R.T., Giedlin, editing. In addition, comprehensive least because gene disruption is the M.A., Nichol, G., et al. (2014). Gene editing of genomic analyses showed no evidence simplest type of gene editing to achieve. CCR5 in autologous CD4 T cells of persons inof adverse or off-target editing events, As noted by Xu et al., its suitability was fected with HIV. N. Engl. J. Med. 370, 901–910. which is a major concern for gene editing recently questioned by a report that Wang, C.X., and Cannon, P.M. (2016). The clinical therapies. That said, the broader applica- CCR5D32 homozygotes have shorter life- applications of genome editing in HIV. Blood 127, 2546–2552. bility of their data does have limitations. spans (Wei and Nielsen, 2019), although The editing protocol used is unlikely to this paper has now been retracted. Xu Wei, X., and Nielsen, R. (2019). Retraction Note: CCR5-D32 is deleterious in the homozygous state be acceptable in the US since CRISPR- et al.’s finding of successful engraftment in humans. Nat. Med., in press. Published online Cas9 reagents were delivered using multi- and persistence of CRISPR-edited October 8, 2019. https://doi.org/10.1038/s41591ple small-scale reactions and research- HSPCs provides a welcome reassurance 019-0637-6. grade equipment. Technical problems of the safety and feasibility of somatic Xu, L., Wang, J., Liu, Y., Xie, L., Su, B., Mou, D., with isolation of the CD34+ cell product cell gene editing and represents impor- Wang, L., Liu, T., Wang, X., and Zhang, B. (2019). CRISPR-Edited Stem Cells in a Patient with HIV also meant that the final edited cell prod- tant progress on the path to developing and Acute Lymphocytic Leukemia. N. Engl. J. uct was diluted with non-edited bulk effective alternatives for HIV+ individuals Med. 381, 1240–1247.
736 Cell Stem Cell 25, December 5, 2019