Upregulation of NKG2D ligands impairs hematopoietic stem cell function in Fanconi anemia José A. Casado, … , Paula Rio, Juan A. Bueren J Clin Invest. 2022;132(15):e142842. https://doi.org/10.1172/JCI142842. Research Article Immunology Stem cells Graphical abstract Find the latest version: https://jci.me/142842/pdf The Journal of Clinical Investigation   RESEARCH ARTICLE Upregulation of NKG2D ligands impairs hematopoietic stem cell function in Fanconi anemia José A. Casado,1,2 Antonio Valeri,1,2 Rebeca Sanchez-Domínguez,1,2 Paula Vela,1,2 Andrea López,1,2 Susana Navarro,1,2 Omaira Alberquilla,1,2 Helmut Hanenberg,3,4 Roser Pujol,2,5 José-Carlos Segovia,1,2 Jordi Minguillón,2,5 Jordi Surrallés,2,5 Cristina Díaz de Heredia,2,6 Julián Sevilla,2,7 Paula Rio,1,2 and Juan A. Bueren1,2 Division of Innovative Therapies, CIEMAT and Advanced Therapies Unit, IIS-Fundación Jimenez Diaz and Autónoma University, Madrid, Spain. 2Biomedical Center for Research on Rare Diseases (CIBERER), 1 Madrid, Spain. 3Department of Pediatrics III, University Children’s Hospital Essen, University of Duisburg-Essen, Essen, Germany. 4Department of Otorhinolaryngology and Head and Neck Surgery, Heinrich Heine University, Duesseldorf, Germany. 5IIB Sant Pau, Hospital de la Santa Creu i Sant Pau, Univ Autònoma de Barcelona, Barcelona Spain. 6Servei d’Oncologia i Hematologia Pediàtrica, Hospital Universitari Vall d’Hebron, Barcelona, Spain. 7Stem Cell Transplant Unit, Hospital Niño Jesús, Madrid, Spain. Fanconi anemia (FA) is the most prevalent inherited bone marrow failure (BMF) syndrome. Nevertheless, the pathophysiological mechanisms of BMF in FA have not been fully elucidated. Since FA cells are defective in DNA repair, we hypothesized that FA hematopoietic stem and progenitor cells (HSPCs) might express DNA damage–associated stress molecules such as natural killer group 2 member D ligands (NKG2D-Ls). These ligands could then interact with the activating NKG2D receptor expressed in cytotoxic NK or CD8+ T cells, which may result in progressive HSPC depletion. Our results indeed demonstrated upregulated levels of NKG2D-Ls in cultured FA fibroblasts and T cells, and these levels were further exacerbated by mitomycin C or formaldehyde. Notably, a high proportion of BM CD34+ HSPCs from patients with FA also expressed increased levels of NKG2D-Ls, which correlated inversely with the percentage of CD34+ cells in BM. Remarkably, the reduced clonogenic potential characteristic of FA HSPCs was improved by blocking NKG2D–NKG2D-L interactions. Moreover, the in vivo blockage of these interactions in a BMF FA mouse model ameliorated the anemia in these animals. Our study demonstrates the involvement of NKG2D–NKG2D-L interactions in FA HSPC functionality, suggesting an unexpected role of the immune system in the progressive BMF that is characteristic of FA. Introduction In an effort to understand the mechanisms involved in FA Fanconi anemia (FA) is a rare inherited disorder mainly charac- BMF, previous studies have shown that FA cells are hypersensi- terized by congenital abnormalities, bone marrow failure (BMF), tive to various cellular stressors, which include oxidative damage and cancer predisposition (1). Currently, allogeneic transplanta- (7, 8), endogenous aldehydes (9, 10), and cytotoxic inflammato- tion of hematopoietic stem and progenitor cells (HSPCs) is the ry cytokines such as TNF-α and IFN-γ, which, furthermore, are preferential therapeutic option for the rescue of BMF in patients present at increased levels in the BM of patients with FA (11, 12). with FA (2). Nevertheless, recent clinical data have shown that Additionally, elevated TGF-β signaling (13) and insufficient viral HSPC gene therapy may constitute an innovative, low-toxicity clearance in FA cells (14) may account for the BMF characteristic therapeutic option for this life-threatening disorder (3). of FA. Despite advances in the understanding of the mechanisms So far, 23 FA genes have been characterized that encode for involved in BMF of patients with FA, further studies are required proteins participating in the FA/BRCA pathway required for the to fully understand the progression of BMF in this disease. repair of DNA interstrand crosslinks (ICLs) and for the preserva- Regarding the involvement of the immune system in BMF dis- tion of genome stability (see reviews in refs. 4, 5). Additionally, it eases, previous studies have indeed implicated cells of the immune has been proposed that at least some FA proteins participate in oth- system as important factors in acquired BMF syndromes (15, 16). er cellular functions, whose alterations can also play a significant However, there is no reported evidence showing the relevance of role in FA disease (see review in ref. 6). specific interactions between immune cells and HSPCs in the pro- gressive BMF that is characteristic of FA. In general, cells of the innate immune system, including NK cells, are generated in BM Conflict of interest: JAC, SN, PR, JCS, JSevilla, and JAB are inventors on patents licensed to Rocket Pharmaceuticals Inc. (Gene therapy for patients with Fanconi and migrate to peripheral lymphoid organs, where they mature anemia. PCT/US2017/050837; Methods for gene modification of hematopoietic cells. and undergo education (17). The cytotoxic function of these cells PCT/US17/264,275). JAB, J Surralles, J Sevilla, and JCS have received consultancy fees depends on the integration of activation and inhibitory signals from from Rocket Pharmaceuticals Inc. an array of surface receptors (17). NKG2D is an activating homod- Copyright: © 2022, Casado et al. This is an open access article published under the imeric member of the NKG2 family of C-type lectin–like receptors terms of the Creative Commons Attribution 4.0 International License. Submitted: August 18, 2020; Accepted: May 25, 2022; Published: August 1, 2022. that enables immune cells to recognize damaged/stressed autolo- Reference information: J Clin Invest. 2022;132(15):e142842. gous cells by interaction with cell-surface ligands indicative of cel- https://doi.org/10.1172/JCI142842. lular distress (18). The human NKG2D receptor is mainly present 1 RESEARCH ARTICLE The Journal of Clinical Investigation   on NK cells, αβ CD8+ T cells, NKT cells, and a subset γδ T cells. of these cells. Studies in complemented and noncomplemented When engaged, NKG2D generates one of the strongest activating fibroblasts from 3 FA-A patients showed that genetic correction with signals in NK cells and also provides a costimulatory signal in CD8+ FANCA retroviral vectors (+FANCA RVs) reduced NKG2D-L levels T cells (19, 20). Ligands of NKG2D (NKG2D-Ls) are 8 membrane when compared with their negative controls (untransduced FA-A proteins MICA, MICB, and ULBP1 to ULBP6, which are general- cells or cells transduced with a noncorrecting FANCG RV (+FANCG) ly absent on healthy cells. However, high expression of NKG2D- (Figure 1C). As expected, the effect of gene complementation on Ls can typically be found in cancer cells and in somatic cells that NKG2D-L levels was even more marked when FA cells were exposed either harbor DNA damage or are affected by oxidative stress, viral to MMC (Figure 1C). Consistent with these studies, enhanced levels infection, or other cellular stressors (21–24). Different studies have of NKG2D-Ls were confirmed in the immunofluorescence analyses shown that overexpression of these ligands enables NKG2D+ effec- conducted in uncorrected FA-A fibroblasts and T cells (Supplemen- tor cells to efficiently recognize and eliminate NKG2D-L+ cells tal Figure 1B). We also observed that, as happened with the expres- (23, 25). NKG2D–NKG2D-L interactions play an important role in sion of NKG2D-Ls, the proportion of cells with nuclear γ-H2AX foci, physiological signaling pathways that recognize stressed or dam- which are indicative of DNA damage, was highest in MMC-treated aged cells, thereby enabling specific cytotoxic cells of the innate uncorrected FA fibroblasts (Supplemental Figure 2). immune system to remove abnormal cells (26, 27). In subsequent studies we evaluated whether NKG2D-L dif- Taking into account the profound phenotypic defects of FA ferences between FA and HD fibroblasts were mainly due to the cells, including a diminished ability to repair DNA crosslinks, upregulated expression of specific ligands. Among the different we hypothesized that these cells might express increased lev- NKG2D-Ls, we found that differences between HD and FA fibro- els of NKG2D-Ls. This upregulation of NKG2-Ls on FA HSPCs blasts were mainly due to the overexpression of MICA, MICB, would promote a deleterious hematopoietic response by immune ULBP-3, and ULBP-4 (Supplemental Figure 3). In contrast to these cells, thus contributing to the progressive BMF observed in most ligands, no significant ULBP-1 upregulation was observed in FA patients. Data presented in this study demonstrates that, com- cells, and in the case of ULBP-2, -5, and, -6, these ligands were pared with healthy cells, cultured fibroblasts and peripheral already upregulated in HD fibroblasts. blood (PB) T cells from patients with FA expressed high levels of Since aldehydes are physiological metabolites with partic- NKG2D-Ls, a process that was further exacerbated by DNA cross- ular damaging effects in FA cells (10), we investigated whether, linker agents. Moreover, a high proportion of fresh CD34+ cells as happened with MMC treatment, formaldehyde also induces from patients with FA also expressed high levels of NKG2D-Ls, NKG2D-L overexpression in FA fibroblasts. As shown in Figure 1D, which correlated with low numbers of CD34+ cells. These data, there were significant differences in the induction of NKG2D-Ls together with in vitro as well as in vivo proof-of-principle exper- in FA fibroblasts treated with 50 μM formaldehyde (see NKG2D-L iments showing improved hematopoietic effects associated with MFI ratios between formaldehyde-treated and untreated FA fibro- the blockage of NKG2D–NKG2D-L interactions, demonstrate the blasts). Representative analyses of NKG2D-L levels (MFI values) relevance of cells of the innate immune system in the impaired in untreated and formaldehyde-treated FA fibroblasts — either growth of FA HSPCs and strongly implicate NKG2D–NKG2D-L uncorrected or gene-corrected — are shown in the right panels of interactions in the BMF characteristic of patients with FA. Figure 1D. These results confirmed that this endogenous genotox- ic agent also resulted in high levels of NKG2D-Ls in FA fibroblasts. Results To analyze the expression levels of NKG2D-Ls in a different FA fibroblasts and T cells express upregulated levels of NKG2D-Ls somatic cell type, we used PB T cells from HDs and patients with in a DNA damage–dependent manner. In order to investigate our FA. As observed in FA fibroblasts, higher basal levels of NKG2D- hypothesis that FA cells might express elevated levels of NKG2D- Ls were detected in in vitro–cultured T cells from patients with FA Ls, we first evaluated the basal expression of NKG2D-Ls in in when compared with basal levels in HD T cells (Figure 2A and Sup- vitro–cultured fibroblasts from patients with mutations in FANCA plemental Figure 1). Additionally, while no increases in NKG2D-L (patients of the FA-A complementation group) and from healthy levels were observed in HD T cells after MMC treatment (MFI donors (HDs). Since all NKG2D-Ls interact with the same NKG2D ratios of MMC-treated cells to untreated cells were close to 1), receptor (28), cells were stained with a mixture of monoclonal this drug significantly increased the levels of NKG2D-Ls in FA T antibodies that recognize the 8 NKG2D-Ls (MICA, MICB, and cells (MFI ratios in cells exposed to 33 nM and 100 nM increased ULBP1–6), as previously described (29). Flow cytometric analy- 2.9- and 4.2-fold, respectively; Figure 2B and Supplemental Figure ses of NKG2D-Ls revealed significant increases in the membrane 1). To examine the effect of gene complementation on NKG2D-L expression of these ligands in FA versus HD fibroblasts (see ratios levels in FA T cells, samples previously transduced with FANCA of the MFI values in Figure 1A and raw MFI data in Supplemen- or FANCG RVs were maintained for 3 days in the absence or the tal Figure 1A; supplemental material available online with this presence of MMC (33 nM). As shown in Figure 2C, we observed article; https://doi.org/10.1172/JCI142842DS1). Additionally, we increased levels of NKG2D-Ls when noncomplemented FA-A T observed that the DNA cross-linker agent mitomycin C (MMC) cells (transduced with the FANCG vector) were exposed to MMC. further increased the differences in NKG2D-L levels between FA However, almost no changes in NKG2D-Ls levels were induced and HD fibroblasts (Figure 1B and Supplemental Figure 1A). by MMC in complemented FA-A cells (Figure 2C). As expected, Next, we investigated whether the high NKG2D-L levels observed NKG2D-L levels observed in HD T cells either transduced with in FA-A cells were a direct consequence of their genetic defect FANCA or FANCG vectors were not modified by exposure to MMC and could therefore be downregulated by the genetic correction (see representative analyses in Figure 2C). 2 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 The Journal of Clinical Investigation   RESEARCH ARTICLE Figure 1. Comparative analysis of NKG2D-L levels in uncorrected and gene-corrected fibroblasts from patients with FA-A and HDs. (A) Primary fibroblasts from HDs (n = 6) and patients with FA-A (n = 8) were cultured at a similar confluence, and NKG2D-L levels were analyzed with a mixture of anti–NKG2D-L antibodies (anti-MICA, -MICB, and –ULBP1-6). Data show the ratio of the MFI of NKG2D-L staining with respect to basal fluorescence. (B) Fold increase of NKG2D-L levels in MMC-treated (+MMC) versus untreated (–MMC) fibroblasts from HDs (n = 7) and patients with FA-A (n = 8). Primary fibroblasts were maintained for 3 days in the absence or presence of MMC (33 nM and 100 nM). NKG2D-L MFI ratios of MMC-treated to nontreated cells are shown. (C) Levels of NKG2D-L (MFI values) in uncorrected and gene-corrected fibroblasts from 3 patients with FA-A. The fibroblasts were exposed to 0 nM, 33 nM, or 100 nM MMC. Corrected cells were generated by transduction with a RV carrying the FANCA gene (+FANCA), whereas uncorrected cells consisted of either untrans- duced cells (MOCK) or cells transduced with a RV encoding FANCG (+FANCG). Representative histograms of NKG2D-L levels in fibroblasts from patient FA-707 are also shown (lower panels). (D) Fold increase of NKG2D-L levels in formaldehyde-treated fibroblasts versus untreated fibroblasts corresponding to FA-competent cells (HD or corrected FA-A cells; FA-A+FANCA) or FA pathway–deficient cells (untransduced; FA-A or FANCC or FANCG-transduced FA-A cells; FA-A+FANCC/G). Data in A, B, and D represent the mean ± SEM. A 2-tailed, unpaired t test was used to compare mean values between HD and FA fibroblasts, either corrected or uncorrected. J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 3 RESEARCH ARTICLE The Journal of Clinical Investigation   Figure 2. Comparative analysis of NKG2D-L levels in uncorrected and gene-corrected T cells from patients with FA-A and HDs. (A) Comparative analysis of NKG2D-Ls in in vitro–cultured T cells from HDs (n = 9) and patients with FA-A (n = 6). Mononuclear PB cells from HDs and patients with FA-A were stimulated with anti-CD3 and anti-CD28 mAbs for 4 days and analyzed for NKG2D-L expression. The figure shows the ratio of the MFI after staining with NKG2D-L antibodies with respect to the basal fluorescence level. (B) Fold increase of NKG2D-L levels in MMC-treated (33 nM or 100 nM) versus untreated T cells from HDs (n = 12, 33 nM and n = 10, 100 nM) and patients with FA-A (n = 8, 33 nM and n = 6, 100 nM). NKG2D-L MFI ratios between MMC-treated and untreated cells are shown. (C) Analysis of MMC-induced (33 nM) expression of NKG2D-L in uncorrected (+FANCG) and corrected (+FANCA) T cells from 3 patients with FA-A. Flow cytometric histograms show representative NKG2D-L analyses in HD and FA-A T cells transduced with the FANCA or FANCG vec- tor. Data in A–C represent the mean ± SEM. In A and B, differences between HD and FA cells were assessed by unpaired t test. In C, differences between corrected and uncorrected samples were evaluated using a paired t test. Taken together, these results demonstrate that, compared FA and HDs, Tables 1 and 2 show the actual percentage of CD34+ with HD cells, primary FA-A fibroblasts and T cells expressed cells in BM, as well as the proportion of CD34+ cells positive for higher levels of NKG2D-Ls. Additionally, our data show that the NKG2D-L expression. expression of these markers of cellular damage/stress was further As expected, the proportion of CD34+ cells in BM from increased after genotoxic damage. patients with FA (0.29% ± 0.08%) was significantly lower than Upregulation of NKG2D-Ls in fresh BM CD34+ cells from patients that determined in BM from HDs (1.98% ± 0.37%) (Tables 1 and with FA. In order to investigate whether the expression of NGK2D- 2). The expression levels of NKG2D-Ls in HD and FA CD34+ Ls was also increased in primary HSPCs from patients with FA, we cells were also investigated by flow cytometry using the gating analyzed fresh BM CD34+ cells from 23 nonmosaic FA patients strategy shown in Supplemental Figure 4A. The results shown in and 7 HDs (Figure 3 and Tables 1 and 2). In some patients (FA- Tables 1 and 2 and in Figure 3, A and B, revealed that the propor- 610, FA-611, and FA-655), we performed repeated BM aspirations, tion of CD34+ cells positive for NKG2D-Ls was markedly high- which allowed us to include a total of 27 FA BM samples in these er in patients with FA (36.55% ± 4.23%) than in HDs (5.48% ± studies. In addition to the main characteristics of patients with 1.34%). Additionally, the levels of NKG2D-Ls were much higher 4 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 The Journal of Clinical Investigation   RESEARCH ARTICLE Table 1. Characteristics of patients with FA included in the study Patient with FA Sample NKG2D-Ls+ in CD34+ cells Age at BM % CD34+ cells >12% Sex FA subtype DEB test BMF No. Ref. No. Ref. sampling (yr) in BM % (highest % in HDs) 1 FA-328 1 FA328 F A 15 68 0.02 43.5 Yes Mild 2 FA-363 2 FA363 F A 7 56 0.06 34.7 Yes Mild 3 FA-535 3 FA535 M A 42 68 0.08 64.8 Yes Severe 4 FA-537 4 FA537 F A 12 64 0.25 24.2 Yes Mild 5 FA-577 5 FA577 F A 6 76 0.62 66.7 Yes Severe 6 FA610 (1) 7 98 0.04 51.8 Yes Mild 6 FA-610 7 FA610 (2) F A 8 ND 0.06 71.7 Yes Mild 8 FA610 (3) 9 ND 0.04 57.9 Yes Severe 9 FA611 (1) 10 74 0.03 57.3 Yes Mild 7 FA-611 M A 10 FA611 (2) 11 ND 0.02 50 Yes Mild 8 FA-616 11 FA616 M A 4 80 0.32 69.1 Yes Severe 9 FA-630 12 FA630 M A 16 84 0.14 23.6 Yes Mild 10 FA-642 13 FA642 M A 5 94 0.57 50 Yes Severe 11 FA-696 14 FA696 M No – A/C/G 5 86 0.08 22 Yes Severe 12 FA-714 15 FA714 F G 13 88 0.01 42.5 Yes Severe 13 FA-771 16 FA771 F G 5 48 0.22 17.9 Yes Severe 14 FA-847 17 FA847 M A 11 94 0.04 46.4 Yes Severe 15 FA-884 18 FA884 F A 9 68 0.2 19.5 Yes No 16 FA-894 19 FA894 F A 8 80 0.4 15.1 Yes Severe 17 FA-344 20 FA344 F D2 14 68 0.06 53.7 Yes Severe 18 FA-652 21 FA652 M A 4 62 0.29 27.5 Yes No 19 FA-830 22 FA830 F A 6 88 0.07 48.8 Yes No 20 FA-649 23 FA649 M A 9 59 0.1 5.3 No Severe 24 FA655 (1) 2 86 1.5 8.2 No No 21 FA-655 M A 25 FA655 (2) 5 ND 0.98 7 No No 22 FA-712 26 FA712 M A 6 66 0.41 3.7 No No 23 FA-739 27 FA739 M A 5 80 1.3 4 No Severe >12% Distribution F/M Mean ± SEM Mean ± SEM Mean ± SEM Mean ± SEM (highest BMF FA subtype % in HDs) Summary of patients with FA 8 mild, 19 A, 2 G, 1 D2, 11 F/12 M 9.4 ± 1.4 75.4 ± 2.7 0.29 ± 0.07 36.5 ± 4.2 22 yes/5 no 13 severe, 1 No – A/C/G 6 No BMF A DEB test was performed to determine the percentage of T cells with aberrant chromosomes. F, female; M, male; ND, not determined. in CD34+ cells from patients with FA than in those from HDs (see 3C, we detected significantly lower numbers of CD34+ cells in the NKG2D-L expression in representative analyses from Figure 3A). group with high expression of NKG2D-Ls. Moreover, when the In 2 patients with FA, we were able to demonstrate an elevated proportion of CD34+ cells positive for NKG2D-Ls was compared proportion of CD34+ cells with high levels of NKG2D-Ls in peri- with the percentage of CD34+ cells in BM, we found an inverse odic BM aspirations spaced 6 months to 2 years apart, revealing correlation (Figure 3D). These observations thus associated high that the expression of NKG2D-Ls in FA HSPCs was not a sporadic expression of NKG2D-Ls with reduced numbers of CD34+ cells in or transient phenomenon (Table 1 and Figure 3B). Interestingly, the BM of patients with FA. analyses of NKG2D-Ls in CD34– and CD34+ cells obtained from To further demonstrate the potential relationship between the the same BM sample reproducibly showed that, in contrast to expression of NKG2D-Ls in CD34+ cells and BMF progression, we CD34+ cells, only a small proportion of CD34– cells had upregulat- identified the patients with a clinical diagnosis of mild or severe ed levels of NKG2D-Ls (Supplemental Figure 4B). BMF (see last column of Table 1). Remarkably, 86.4% of FA sam- We next investigated whether expression of NKG2D-Ls ples (19 of 22) included in the high NKG2D-L expression group affects the proportion of CD34+ cells in BM from patients with FA. were associated with a patient diagnosis of BMF. Since the highest proportion of HD BM CD34+ cells that expressed Taken together, these results demonstrate that, in contrast to NKG2D-Ls was 12% (Tables 1 and 2), we classified FA BM samples HD CD34+ cells, a high proportion of BM CD34+ cells from patients into groups above and below this threshold. As shown in Figure with FA overexpressed NKG2D-Ls. Additionally, these data indicate J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 5 RESEARCH ARTICLE The Journal of Clinical Investigation   increase in the proportion of NK cells Table 2. Characteristics of HDs included in the study (CD56+CD3–) was observed in BM sam- NG2D-L+ in CD34+ cells ples from patients with FA compared FA Age at BM % CD34­+ cells >12% with samples from HDs. Differences HD no. Sex DEB test BMF subtype sampling (yr) in BM % (highest became significant for NK cells with a % in HDs) high capacity to mediate natural and HD1 F NA 19 NA 1.36 5.1 No No antibody-dependent cellular cytotox- HD2 M NA 2 NA 1.41 12 No No icity (CD56+loCD16+hiCD3–) (32), and HD3 M NA 7 NA 1.40 1.6 No No also for activated and tissue-resident HD4 M NA 2 NA 3.84 5.6 No No NK cells (CD56+CD69+CD3–) (33). We HD5 M NA 19 NA 1.8 7.1 No No also detected significantly higher per- HD6 F NA 15 NA 1.2 1.6 No No centages of both CD8+CD3+ and acti- HD7 M NA 21 NA 2.9 5.4 No No vated CD8+CD69+CD3+ cytotoxic T F/m Mean ± SEM Mean ± SEM Mean ± SEM Summary of HDs NA NA 7 No 7 No cells in BM samples from patients with 2F/5M 12.1 ± 3.1 1.98 ± 0.37 5.48 ± 1.3 FA compared with those from HDs, whereas no differences were observed for CD4+CD3+ cells (Figure 5A). Final- ly, we confirmed the similar expression that overexpression of NKG2D-Ls was associated with both reduced levels of the NKG2D receptor on CD56+ T cells as well as on CD8+ numbers of CD34+ cells in BM and a high incidence of BMF. T cells from patients with FA and HDs (Figure 5B). The downregulation of FANCA in HD CD34+ cells increases These results demonstrated an increased percentage of activat- NKG2D-L levels in a CHEK1-dependent manner. To prove the rele- ed NKG2D+ cytotoxic immune cells in BM from patients with FA, vance of a functional FA/BRCA pathway for the HSPC NKG2D-L which might generate detrimental effects in NKG2D-L–expressing levels, CD34+ cells from HD cord blood (CB) samples were trans- HSPCs residing in this tissue. duced with lentiviral vectors that expressed either an anti-FANCA Inhibition of the clonogenic potential of FA patients’ HSPCs medi- shRNA together with an EGFP marker gene (shFANCA/EGFP LV; ated by NKG2D–NKG2D-L interactions. In this set of experiments, see Methods) or a scrambled shRNA and EGFP (shSCR/EGFP we determined whether FA CD34+ cells expressing NKG2D-Ls LV). Four to 12 days after transduction, we analyzed the levels of consisted of cells already destined to die, even in the absence NKG2D-Ls in EGFP+ cells by flow cytometry and detected high- of interactions with immune cells. First, analyses of annexin V er levels of NKG2D-Ls in CD34+ cells that had been transduced expression in BM samples from 4 patients with FA revealed that with the shFANCA/EGFP LV (FA-like CD34+ cells) compared with 43%–73% of FA CD34+NKG2D-L+ cells were annexin V–, indi- samples transduced with the shSCR/EGFP LV (see individual data cating that a high proportion of these cells did not show signs of and representative analyses in Figure 4A). Similarly, when ratios apoptosis (Supplemental Figure 6). To investigate the prolifer- of NKG2D-L MFIs between EGFP+ and EGFP– cells from the same ation integrity of CD34+ cells that expressed NKG2D-Ls, in sub- cultures were compared, we observed significantly higher ratios in sequent experiments, we determined the clonogenic potential of the FA-like cells (MFI ratio: 1.41) compared with ratios in the con- purified NKG2D-L+ and NKG2D-L– CD34+ cells from 2 patients trol group (MFI ratio: 1.07) (Figure 4B). with FA. Although sorted NKG2D-L– cells from both patients Since it has been previously reported that the DNA damage generated higher colony numbers compared with NKG2D-L+ response (DDR) can lead to increased levels of NKG2D-Ls in a pro- cells, we found that a number of colonies were also generated cess dependent on ataxia telangiectasia and Rad3-related protein/ by NKG2D-L+ cells (range: 42%–81.9% of colony numbers gen- checkpoint kinase 1 (ATR/CHEK-1) activation (30, 31), we ana- erated by NKG2D-L– cells; Figure 6A). These results thus con- lyzed whether the upregulation of NKG2D-Ls in FA-like HSPCs firmed the proliferation integrity of an important number of was dependent on ATR/CHEK-1 activation. With this aim in mind, NKG2D-L-expressing FA CD34+ cells. we conducted 2 experiments in which FA-like CD34 cells were + To evaluate the effect of NKG2D–NKG2D-L interactions incubated with a specific inhibitor of CHEK-1 (SB-2118078) that between cells of the immune system (such as NK and CD8+ T prevents the MMC-mediated accumulation of FA cells in the G2/M cells) and HSPCs, we conducted in vitro NKG2D–blocking assays phase of the cell cycle and also the phosphorylation of CHEK-1 in with BM samples from patients with FA and HDs. In these stud- FA cells (Supplemental Figure 5, A and B, respectively). Remark- ies, mononuclear BM cells from each donor, containing both ably, the CHEK-1 inhibitor downregulated NKG2D-L levels in the immune effector cells and the target HSPCs, were incubat- FA-like CD34+ cells (Figure 4C), further supporting the implication ed either with an anti-NKG2D–blocking antibody (a-NKG2D) or of DDR in the upregulation of NKG2D-Ls in FA HSPCS with a matching isotype control. Thereafter, we performed clono- An increased proportion of NK and CD8+ T cells is observed in BM genic assays to assess the effects mediated by NKG2D receptor from patients with FA. Since NK and T cells equipped with NKG2D blockage on colony numbers (see the experimental protocol in Fig- receptors could mediate deleterious effects in NKG2D-L–express- ure 6B). In the case of HD samples, colony numbers generated in ing HSPCs, we first determined the proportion and activation the presence of the a-NKG2D mAb and the isotype control were phenotype of NK and CD8+ T cells in BM from HDs and patients very similar, with a ratio of colony numbers of 0.98 ± 0.02. Never- with FA. As shown in Figure 5A, a moderate though nonsignificant theless, when FA BM samples were tested in the same manner, this 6 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 The Journal of Clinical Investigation   RESEARCH ARTICLE Figure 3. Analysis of NKG2D-L levels in fresh BM CD34+ cells from HDs and patients with FA. (A) Representative contour plot analyses of NKG2D-L in CD34+ cells from HDs (n = 3) and patients with FA (n = 3). SSC-A, side scatter area. (B) Proportion of CD34+ cells positive for NKG2D-L in BM from HDs (n = 7) and patients with FA (n = 23). Light blue data points correspond to FA BM samples from 1 single analysis (n = 20). Dark blue (FA-610), green (FA-611), and gray (FA-655) data points correspond to BM samples from the same patient with FA analyzed at different time points (n = 27 total FA samples). Additional details on the FA and HD samples are provided in Tables 1 and 2. (C) Analysis of the percentage of CD34+ cells in BM samples from HDs and patients with FA. FA samples were classified into 2 groups according to NKG2D-L levels, using as a cutoff the maximum percentage of NKG2D-Ls determined in HD samples (12%; see Tables 1 and 2). (HD, n = 7; FA >12% NKG2D-L+, n = 22; FA ≤12% NKG2D-L+, n = 5). P values were determined by Kruskal-Wallis test followed by Dunn’s mul- tiple-comparison test. (D) Correlation analysis between the percentage of CD34+ cells in BM and the proportion of CD34+ cells positive for NKG2D-Ls in samples from HDs and patients with FA (data point colors as in B, n = 27) and HDs (red data points, n = 7). Mean values ± SEM are indicated for each group in B and C, and significance was calculated by unpaired t test. Spearman’s correlation test was applied for assessment of HD and FA samples and also only for FA samples. Dashed lines indicate the lowest percentage of CD34+ cells and the highest proportion of NKG2D-L expression in BM CD34+ cells from HDs. ratio was significantly increased to 1.31 ± 0.04 (P < 0.001) (see ratios Blockage of NKG2D–NKG2D-L interactions ameliorates anemia in Figure 6B and total colony numbers in Supplemental Figure 7). in a BMF FA mouse model. In a final set of experiments in a BMF Since NKG2D is a key activation receptor for NK cells, we FA mouse model, we sought to determine whether the blockage conducted additional studies with purified CD56+ cells to eval- of NKG2D–NKG2D-L interactions is associated with protective uate the inhibitory effects of these cells on purified CD34+ cells hematopoietic effects. Because Fanca–/– mice do not develop obtained from the same HD or FA BM sample. Similar to the effects spontaneous BMF, we treated these mice with MMC as described observed with nonpurified samples, we found that the numbers of previously for Fancc–/– mice (34). Concomitant to the administra- colonies were increased when the a-NKG2D mAb was added to the tion of MMC, mice were treated with a blocking a-NKG2D mAb cocultures of purified CD34+ and CD56+ cells from patients with or with an isotype control (Figure 7A). As shown in representative FA. Importantly, increases in the number of colonies did not occur analyses from Figure 7, B and C, MMC markedly increased the when purified samples from HDs were used (Figure 6B and Sup- proportion of hematopoietic progenitors (Lin– c-Kit+) and prim- plemental Figure 7). Taken together, these in vitro experiments itive hematopoietic stem cells (HSCs) (Lin– Sca1+c-Kit+ HSCs; demonstrate that immune cells present in the BM of patients with LSK cells) expressing NKG2D-Ls. Administration of the a-NK- FA, including purified NK cells, induced an NKG2D-mediated G2D mAb further increased the proportion of HSPCs expressing inhibitory effect against autologous HSPCs. NKG2D-Ls, although the differences did not reach statistical J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 7 RESEARCH ARTICLE The Journal of Clinical Investigation   Figure 4. Effects of FANCA knockdown on NKG2D-L levels in HD CD34+ cells. (A) Comparative analysis of NKG2D-L in HD CB CD34+ cells transduced with a control shRNA (shSCR/EGFP) or FANCA shRNA (shFANCA/EGFP) LVs. Left panel: NKG2D-L levels in EGFP+CD34+ cells that were transduced with either LV for 4–12 days prior to NKG2D-L analyses (n = 9). Right panels: Representative histograms of NKG2D-L levels in CD34+EGFP+ cells transduced with either LV. NKG2D-Ls were quantified 7 days (Exp 1) or 12 days (Exp 2) after transduction. (B) Left panel shows comparative analyses of NKG2D-L MFI ratios between EGFP+ and EGFP– cells in samples from the same culture as in A (n = 9). Right panels: Representative histograms showing comparative levels of NKG2D-L in EGFP+ versus EGFP– cells after transduction with shSCR/EGFP or shFANCA/EGFP LVs. (C) Seven days after transduction, FA-like CD34+ cells were incubated with a CHEK-1 inhibitor or DMSO vehicle, and NKG2D-L levels were analyzed in EGFP+ cells. Histograms show NKG2D-L levels in 2 inde- pendent experiments after incubation with a CHEK-1 inhibitor for 24 hours (Exp 1) or 48 hours (Exp 2). Differences in A were determined by paired t test. Differences in B were determined by Mann-Whitney U test. Mean values ± SEM are shown. significance. In all experimental conditions, we confirmed the CD8+ and CD4+ T cells in MMC-treated mice was similar in the presence of NK cells and CD8+ and CD4+ T cells in BM from Fan- a-NKG2D and isotype treatment groups. ca–/– mice (Figure 7D). In these analyses, we also observed that, When hematological parameters were analyzed (Figure 7E), although the overall proportion of NK cells in BM samples was we observed higher hemoglobin values and numbers of RBCs decreased after MMC treatment, the percentage of NK cells and in Fanca–/– mice treated with the a-NKG2D mAb compared with 8 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 The Journal of Clinical Investigation   RESEARCH ARTICLE Figure 5. Comparative analysis of BM NK and T cells from HDs and patients with FA. (A) Percentage of NK cells (CD56+CD3–; FA n = 22, HD n = 8) and of NK cells with strong cytotoxic capacity (CD56+/loCD16+CD3–; FA n = 15, HD n = 6) and activated NK cells (CD56+CD69+CD3–; FA n = 14, HD n = 6) in mononuclear BM cells from HDs and patients with FA. Lower panels show the corresponding percentages of BM cytotoxic T cells (CD8+CD3+; FA n = 11, HD n = 8), acti- vated cytotoxic T cells (CD8+CD69+CD3+; FA n = 11, HD n = 6), and Th cells (CD4+CD3+; FA n = 6, HD n = 9). To compare data between HD and FA samples, an unpaired t test was used for comparisons between CD8+CD3+ cells, with Welch’s correction for CD56+loCD16+CD3– and CD4+CD3+ cells, and the Mann-Whit- ney U test was used for comparisons between CD56+CD3–, CD56+CD69+CD3–, and CD8+CD69+CD3+ cells. (B) Left panels show representative histograms of NKG2D receptor levels in NK and CD8+ T cells from HDs and patients with FA. Right panels show comparative analyses of NKG2D receptor expression in BM samples from HDs and patients with FA (n = 3). An unpaired t test was used for comparisons of NKG2D receptor expression between HD and FA samples in CD56+ and CD8+ cell populations. Mean values ± SEM are shown in all panels. isotype-treated control mice. These values reached statistical Discussion significance on day 14 after MMC treatment. In contrast, no dif- Previous studies have shown that DNA damage generated by exog- ferences between the 2 groups of mice were observed when the enous or endogenous toxic sources, and even the replicative stress numbers of WBCs and platelets were considered. of FA cells, can induce DDR signaling, which activates ATR and Overall, this final set of in vivo experiments demonstrates ataxia telangiectasia mutated (ATM) cellular sensors as well as that the blockage of NKG2D–NKG2D-L interactions mediated a downstream proteins such as CHEK-1 and TP53 (35, 36). Given pre- protective effect in the erythrocyte lineage in a BMF FA mouse vious evidence showing that the DDR upregulates the expression model. of NKG2D-Ls, and taking into account that NKG2D-L–expressing J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 9 RESEARCH ARTICLE The Journal of Clinical Investigation   Figure 6. Clonogenic potential of NKG2D-L+ and NKG2D-L– FA CD34+ cells and NKG2D receptor blockade effects. (A) Analysis of the clonogenic poten- tial of NKG2D-L–expressing and nonexpressing BM CD34+ cells from 2 patients with FA. Left panels show the cell-sorting strategy used to determine the clonogenic potential of NKG2D-L+ and NKG2D-L– CD34+ cell fractions. FMO corresponds to the PE fluorochrome used for the sorting of NKG2D-L+ and NKG2D-L– cells. Right graphs show the colony numbers generated by NKG2D-L+ and NKG2D-L– CD34+ cells from patients with FA. The numbers above the red bars indicate the percentage of colonies generated by NKG2D-L+ compared with NKG2D-L– CD34+ cells. (B) Effects mediated by NKG2D receptor block- ade on the clonogenic potential of BM samples from HDs and patients with FA. Left panel shows the experimental protocol used to evaluate the clonogen- ic effects mediated by inhibition of the NKG2D receptor. Right panel shows the ratio between the number of hematopoietic colonies generated by HD (n = 9) or FA (n = 23) BM samples after incubation with an anti-NKG2D–blocking antibody (a-NKG2D) or an isotype control. Black data points represent colony numbers generated in experiments performed with mononuclear BM cells. Some experiments (red data points) were performed using purified BM CD34+ cells (as target cells) and purified BM NK cells (as effector cells) from the same donor (effector/target cell ratio of 10:1). Mean values ± SEM are shown. An unpaired t test was used to compare mean values. Total colony numbers are shown in Supplemental Figure 7. cells could be recognized by cells of the immune system (25, 27, 30, prevented the upregulation of these molecules, showing the rele- 31), we hypothesized that FA cells could be prone to overexpress vance of this DNA repair pathway in the modulation of NKG2D-Ls. NKG2D-Ls, and also that this upregulation on FA HSCs could play a Although all NKG2D-Ls interact with the same NKG2D recep- relevant role in the etiology of BMF progression in patients with FA. tor (28), we observed that increased levels of NKG2D-Ls in FA Our first studies in untreated fibroblasts and PB T cells showed fibroblasts were mainly due to elevated levels of MICA as well that NKG2D-L levels were significantly higher in FA cells than in HD as of MICB, ULBP3, and ULBP4. Because of the complexity of cells, and that either MMC or formaldehyde — a physiological source the regulatory mechanisms involved in modulating NKG2D-Ls, inducing endogenous DNA damage (37) — significantly increases including transcriptional, posttranscriptional, and posttransla- NKG2D-L levels in FA cells. The direct involvement of the FA path- tional levels (see review in ref. 38), specific studies should be per- way in the regulation of NKG2D-L levels was confirmed in experi- formed to clarify the mechanisms accounting for the upregulated ments in which the mere complementation of the FA genetic defect expression of the individual NKG2D-Ls in different FA cell types. 10 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 The Journal of Clinical Investigation   RESEARCH ARTICLE Figure 7. Implications of NKG2D–NKG2D-L interactions in a Fanca–/– mouse BMF model. (A) Experimental protocol used for evaluating the effect of NKG2D– NKG2D-L interactions in the BMF induced by MMC in Fanca–/– mice. Mice were periodically treated with an isotype control or an anti-NKG2D mAb, prior to and after initiating MMC treatment (2 doses of 0.3 mg/kg). PB samples from Fanca–/– mice were obtained prior to and after treatment. (B) Representative flow cytometric analysis showing expression of NKG2D-Ls in Lin–c-Kit+ cells from Fanca–/– mice treated with MMC compared with expression in untreated mice. (C) Proportion of Lin–c-Kit+ cells and LSK cells positive for NKG2D-Ls at the end of the experimental protocol (day 14) in untreated or MMC-treated Fanca–/– mice receiving the isotype control or the anti-NKG2D antibody. (D) Proportion of NK cells and CD8+ and CD4+ T cells, and their respective activated subpopula- tions in mononuclear BM cells from Fanca–/– mice treated as in C. (E) Comparative analysis of PB cell parameters in MMC-treated Fanca–/– mice that received the isotype control (blue dots) or the anti-NKG2D mAB (red dots). Each dot shows changes in PB parameters on days 7 and 14 after MMC treatment with respect to the values determined on day –4 (prior to starting mAb and MMC treatments), which were considered as 100%. The mean PB values corresponding to the isotype control– and the anti-NKG2D–treated groups on day –4 were, respectively, hemoglobin: 14.53 and 13.38 g/dL; RBCs: 10.37 and 9.32 × 106 cells/μL; WBCs: 4.64 and 5.08 × 103 cells/μL; platelets: 1002 and 914 × 103 cells/μL. The Mann-Whitney U test was used to compare day 7 mean values and an unpaired t test to compare day 14 mean values according to the respective data distributions. Whiskers represent the minimum and maximum values, the lower and upper box edges correspond to the 25th and 75th percentiles, respectively, and the lines within the boxes correspond to the median. Remarkably, analyses conducted in fresh BM samples revealed from the same BM sample, suggesting that, as shown in different that, in contrast to HD CD34+ cells, a large proportion of CD34+ cells cancer cell types (39), the upregulation of NKG2D-Ls is most evi- from patients with FA expressed high levels of NKG2D-Ls. Never- dent in cells of the stem cell compartment, in this particular case, theless, this upregulation was not observed in CD34– cells obtained in FA HSPCs. Even more relevant was our observation of an inverse J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 11 RESEARCH ARTICLE The Journal of Clinical Investigation   correlation between the proportion of CD34+ cells in FA patients’ erythrocyte lineage in which we first noted the improvement as a BM and the proportion of CD34+ cells expressing NKG2D-Ls, which result of HSC gene therapy in patients with FA-A (3). strongly points to the negative impact of NKG2D-L expression in FA Currently, suppressive NKG2D-mediated processes against HSPCs. These results, together with the observation that a high per- HSPCs have only been associated with the BMF that takes place centage of patients with FA with NKG2D-L–overexpressing CD34+ in patients with acquired aplastic anemias (15, 16). Our study pro- cells showed clinical signs of BMF, further support the relevance of vides evidence of HSPC-inhibitory effects mediated by NKG2D– NKG2D-L expression in the pathogenesis of BMF in FA. NKG2D-L interactions in an inherited BMF syndrome, in particular Our observations showing that the downregulation of FANCA in FA. We propose that the DDR generated by exogenous or endoge- by lentiviral shRNA vectors (FA-like cells) upregulated NKG2D-Ls nous damaging agents and/or by the replicative stress of FA HSPCs levels in HD CD34+ CB cells demonstrate the direct impact of the would result in different cellular alterations leading to NKG2D-L FA/BRCA pathway on NKG2D-L levels in FA HSPCs. Additional- overexpression in these cells. Once recognized by NKG2D+ effector ly, the fact that inhibition of CHEK-1 downregulated NKG2D-Ls immune cells, these would exert cytotoxic effects through perforins in FA-like CD34+ cells strongly supports the hypothesis that the and granzymes and/or the secretion of proinflammatory cytokines involvement of the DDR in increased levels of NKG2D-Ls is a direct such as TNF-α and IFN-γ. These immune mechanisms would thus consequence of the disruption of the FA/BRCA pathway. Interest- contribute to the onset or amplification of an inflammatory environ- ingly, a previous study described an attenuated clinical phenotype ment and increased HSPC damage in FA BM. in patients with FA with mild BMF that was associated with low lev- Overall, our study suggests an unexpected role of the immune els of CHEK-1 in cells of the hematopoietic system (35). As CHEK-1 system in the progressive BMF that is characteristic of patients with inhibition significantly reduced NKG2D-L levels in FA HSPCs, we FA. Detailed follow-up studies in large series of patients with FA can speculate that the low levels of CHEK-1 present in these patients would indicate whether the overexpression of NKG2D-Ls in HSPCs with attenuated FA may prevent the overexpression of NKG2D-Ls, could be used as an early event predicting BMF in patients with FA. limiting the recognition of these primitive hematopoietic cells by Additionally, our results present alternative therapeutic options to immune cells. This process would account, at least in part, for the delay BMF onset in patients with FA, based on the pharmacological improved hematopoiesis observed in these patients. inhibition of NKG2D–NKG2D-Ls interactions. Nevertheless, taking Although CD8+ T cells and NK cells mainly reside in differ- into account the cancer-prone nature of FA cells and the paramount ent lympho-hematopoietic tissues, both cell types can migrate to role of these interactions in the elimination of tumor cells, these inflamed tissues (33, 40, 41). We therefore reasoned that the percent- strategies would need to be carefully evaluated in patients with FA. age and activation status of these NKG2D+ cells might be increased in the BM of patients with FA, a hypothesis that was confirmed in our Methods comparative BM immune-phenotypic studies of HDs and patients Samples from patients with FA and HDs. All patients with FA had a posi- with FA. Whether CD8+ T cells and NK cells were activated in situ or tive chromosomal breakage test, and most of them corresponded to the recruited as already preactivated cells is currently unknown. FA-A complementation group (Table 1). Patients with FA with somatic Although it could be speculated that NKG2D-L–expressing mosaicism were excluded from this study. In all instances, fresh BM HSPCs are already destined to die, our results indicate that this is samples were analyzed within the first 48 hours after BM aspiration. not the case for a significant proportion of these cells. This conclu- Patients with BMF at the moment of the analysis were defined as those sion proceeds from our experiments showing that a proportion of with hypocellular BM, and 2 of the following PB criteria: hemoglobin FA CD34+ NKG2D-L+ cells are characterized by negative annexin <10 g/dL; neutrophils <1.5 × 109/L; platelets <50 × 109/L. Patients with V staining and retain a significant clonogenic capacity. Moreover, only 1 affected PB lineage were defined as having mild BMF. The per- NKG2D-blocking experiments with either purified or nonpurified centage of aberrant T cells in the diepoxybutane (DEB) breakage test CD34+ and NK cells demonstrated that, at least in vitro, immune corresponds to analyses performed at diagnosis. Umbilical CB samples BM cells from patients with FA exert NKG2D-mediated inhibitory to obtain HD CD34+ cells were provided by the Centro de Transfusión effects on the clonogenic potential of FA HSPCs, an effect that was de la Comunidad de Madrid. not observed when HD BM cells were used. Cultures of human fibroblasts and PB T cells. All primary cells were Since NKG2D not only constitutes a strong activating receptor generated at the CIEMAT (Madrid, Spain). Primary skin fibroblasts were of NK cells, but is also a coreceptor in CD8+ cells (19), we cannot rule cultured in DMEM plus GlutaMAX (Gibco, Thermo Fisher Scientific) out the possibility that FA CD8+ cells might also exert some addi- supplemented with 20% FBS (HyClone). Mononuclear PB cells were tional effect against FA HSPCs. However, it seems unlikely that T obtained after Ficoll-Paque PLUS (GE Healthcare) gradient centrifuga- cells are involved in this process, as multiple studies of T cell immu- tion of PB samples. Cells were washed in PBS and cultured in nontreated nosuppression have only shown hematologic improvements in plates coated with purified anti–human CD3 (catalog 555728) and puri- acquired BMF syndromes, but not in inherited ones such as FA (42). fied anti–human CD28 (555356) mAbs (both from BD Biosciences) with In our final set of experiments, we investigated the hemato- IMDM plus GlutaMAX (Gibco, Thermo Fisher Scientific) with 20% FBS poietic effects that NKG2D–NKG2D-L interactions might have in as previously reported to activate T cell proliferation (43). vivo. In these studies, we used a BMF Fanca–/– mouse model based Hydroxyethyl starch sedimentation of BM samples and purification on MMC administration, as previously described in Fancc –/– mice of CD34+ cells from CB. Fresh heparinized BM aspirates from HDs and (34). Remarkably, our data showed that the inhibition of NKG2D– patients with FA were subjected to hydroxyethyl starch sedimentation NKG2D-L interactions enhanced hemoglobin values and RBC (HES) to remove erythrocytes and maximize the number of collected numbers in this BMF mouse model. Strikingly, it was also in the leukocytes (44). Briefly, BM samples were mixed with 6% HES (Grifols) 12 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 The Journal of Clinical Investigation   RESEARCH ARTICLE and left undisturbed for 4 hours at room temperature, allowing RBCs in polystyrene chambers (Culture Slides, Falcon) and incubated for 24 to sediment by gravity. Cells were washed in PBS, and the remaining hours. Thereafter, the medium was replaced by 200 μL medium contain- RBCs were lysed with ammonium chloride solution (8.3 g/L ClNH4, 1 ing the NKG2D-L mAb mixture and incubated for an additional 30 min- g/L KHCO3, and 2 mL 0.5 M EDTA). After 3 minutes, cells were again utes at 37°C. After 2 washing steps with PBS, cells were examined using washed in PBS and resuspended in PBA (PBS and 0.1% BSA). CD34+ an Eclipse Ts 2R microscope (Nikon; ×20 objective, ×0.5 magnification). cells from HD CB were isolated after Ficoll-Paque PLUS gradient centrif- Representative images were obtained with a DFC 7000 T camera (Leica). ugation and subsequent immunomagnetic cell sorting (microBead kit, To test the levels of DNA damage in uncorrected and gene-correct- human, Miltenyi Biotec) according to the manufacturer’s instructions. ed FA fibroblasts, cells were cultured in CultureSlides with MMC (100 Flow cytometry and sorting of cells from human BM samples. To analyze nM). Twenty-four hours after treatment, cells were fixed in 4% para- the expression of all known NKG2D-Ls in HD and FA cells, a mixture of formaldehyde in PBS for 15 minutes at room temperature (RT), and the anti–human NKG2D-L antibodies MICA-PE (clone, 159227), MICB- after 3 washes in PBS, cells were permeabilized in 0.5% Triton X-100 PE (clone, 236511), ULBP1-PE (clone, 170818), ULBP2,5,6-PE (clone, (MilliporeSigma) for 5 minutes. After 3 washes, slides were blocked in 165903), ULBP3-PE (clone, 166510), and ULBP4-PE (clone, 709116) TBS 0.1% NP-40 plus 10% FBS for 30 minutes and incubated with a (all from R&D Systems) was prepared. Analyses of individual NKG2D- primary antibody (anti–phosphorylated histone H2AX [Ser139], clone Ls in fibroblasts were also performed. To analyze the expression of JBW301, MerkMillipore) diluted in the blocking buffer (1 μg/mL) at 4°C NKG2D-Ls in HSPCs, cells were stained with an anti–CD34-FITC anti- overnight. After 3 washes in TBS, cells were incubated for 1 hour at RT body (clone 8G12) from BD Biosciences, together with the mixture of with the secondary fluorescent polyclonal antibody Alexa Fluor 488 NKG2D-L antibodies. Prior to cell labeling with the corresponding anti- F(ab′)2 goat anti–mouse IgG (H+L) (A-11017) from Invitrogen (Ther- bodies, BM cells were incubated with an FcR-blocking reagent (Miltenyi mo Fisher Scientific) diluted with the blocking buffer and DAPI. Sam- Biotec) to prevent nonspecific binding. At least 500,000 mononu- ples were visualized with an Axio Imager.M2 fluorescence microscope clear cells from BM were analyzed per sample to quantify the level of (Carl Zeiss; ×63 objective, ×1.4 magnification). Two hundred cells were NKG2D-Ls in CD34+ cells. CD34-APC (clone 581) from BD Biosciences counted, and cells with more than 10 foci per cell were scored as posi- and Annexin V-FITC/IP from BD Pharmingen were used to analyze tive. Each data point in the figures represents data from 1 independent early apoptosis in CD34+ cells labeled with the mixture of NKG2D- experiment, with the mean of 3 independent experiments also shown. L-PE antibodies. The percentage of CD34+ cells in BM aspirates was Images were captured with an Axiocam 503 mono camera (Carl Zeiss) quantified following the International Society of Hematotherapy and and processed with ZEN 2.3 Blue Edition software (Carl Zeiss). Graft Engineering (ISHAGE) guidelines for CD34+ cell determination CHEK-1 inhibition and cell-cycle studies. Fibroblasts were cultured by flow cytometry using CD34-PE (clone 8G12) from BD Biosciences with 33 nM and 100 nM MMC for 3 days in the presence or absence of and CD45-FITC (clone J33) from Immunotech. To study the immune the CHEK-1 inhibitor SB-2118078 (Calbiochem) at 1 μM. After incubation phenotype in BM, cells were stained with antibodies against the fol- with MMC, cells were washed with PBS and fixed with 70% cold ethanol lowing surface markers: CD3-APC (clone UCHT1) from BioLegend, for 24 hours. Cells were washed again with PBS and incubated for 30 min- CD56-FITC (clone NCAM16.2) from BD Biosciences, CD8-FITC (clone utes at 37°C with 50 μg/mL propidium iodide (MilliporeSigma) and 100 B9.11) from Beckman Coulter, CD4-FITC (clone 13B8.2) from Beckman μg/mL RNase A/T1 (Thermo Fisher Scientific). DNA content was ana- Coulter, CD69-PE (clone FN50) from BioLegend, and CD16-PCy5 lyzed by flow cytometry, avoiding cell aggregates in linear amplification. (clone 3G8) from BD Biosciences. In all flow cytometric studies, DAPI Transduction of human fibroblasts and T cells with RVs. FA-A fibro- was used to discard dead cells in the analysis. Flow cytometry was per- blasts were transduced with RVs carrying the FANCA and Neor genes formed on an LSR Fortessa cytometer (BD Biosciences), and data were (FAIN vector) to generate corrected FA-A cells. As controls, untrans- analyzed using FlowJo software, version 10.8.1 (Tree Star). duced or FA-A cells transduced with RVs carrying the FANCC or Antibodies CD56-PE (clone N90) form Beckman Coulter and FANCG genes, together with the Neor gene (FCIN and FGIN vectors, CD34-APC (clone 581) from BD Biosciences were used for the sort- respectively), were used (45). Transduced cells were selected with ing of BM NK cell populations and HSPCs, respectively. CD34-APC G418 at 0.7 mg/mL (Calbiochem) for 5 days and expanded prior to (clone 581) and CD45-FITC (clone J33) from Beckman Coulter, and use. PB T cells from HDs or patients with FA-A were transduced with the mixture of NKG2D-Ls–PE antibodies were used to discriminate RVs carrying the FANCA and EGFP genes (LFAPEG) or the FANCG BM CD34+NKG2D-L­+ and CD34+NKG2D-L­– cells in a BD Influx Cell and EGFP genes (LGEG11) (43, 46). For the transduction of T cells, Sorter. A fluorescence minus one (FMO) control stained with anti– mononuclear PB cells were preincubated with immobilized anti-CD3 CD34-APC and anti–CD45-FITC antibodies was used to set the gat- and anti-CD28 antibodies (see above) in 6-well, nontreated plates for ing of CD34+NKG2D-L­+ and CD34+NKG2D-L­– cells. Sorted cells were 4 days. Thereafter, the cells were transduced with 2 infection cycles of cultured in methylcellulose-based medium (see below). the corresponding RVs (24 hours per cycle). Colony-forming cell assays. Clonogenic assays were performed MMC and formaldehyde treatments of human fibroblasts and T cells. using methylcellulose-based medium (MethoCult H4434, STEM- Fibroblasts from HDs and patients with FA-A were cultured for 3 days in CELL Technologies). Fourteen days after seeding of the cells, hema- the absence or presence of MMC (33 nM and 100 nM; MilliporeSigma). topoietic colonies (total colonies and CFU-GM and BFU-E colonies) For formaldehyde treatments, fibroblasts were incubated for 24 hours were scored under an inverted phase-contrast microscope at ×40 with 25 μM and 50 μM formaldehyde (prepared from a 37% solution, (Nikon Diaphot). An appropriate number of total or purified BM cells proteomics grade; VWR Life Science). After incubation with MMC were seeded to facilitate the scoring of 10–100 colonies per plate. or formaldehyde, cells were trypsinized and stained to measure total NKG2D-Ls and phosphorylated H2AX immunofluorescence studies. expression of NKG2D-Ls. Activated T cells (see above) were challenged For the NKG2D-L staining of HD and FA fibroblasts, cells were seeded with MMC (33 nM and 100 nM) for 4 days, and NKG2D-Ls expression J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 13 RESEARCH ARTICLE The Journal of Clinical Investigation   in DAPI– cells was analyzed by flow cytometry. To evaluate the basal collected, resuspended, and seeded in methylcellulose cultures (see expression of NKG2D-Ls, the ratios of MFI values between unstained experimental scheme in Figure 6). Colony numbers generated in the cells (–mAb) and NKG2D-L–labeled cells (+mAb) were determined. NKG2D-blocking group were normalized with respect to numbers cor- FANCA interference in HD CD CD34+ cells. FA-like CD34+ cells were responding to the isotype control group. Some experiments were con- generated by transducing HD CB CD34+ cells with a shFANCA/EGFP ducted using purified CD56+ cells (effectors cells) and purified CD34+ LV (47). This LV carries the sh11 and sh7 hairpin RNAs (shRNAs) that cells (target cells) sorted by FACS (see above). In these experiments, a target human FANCA and expresses the EGFP marker gene. The shFAN- ratio of 10:1 (effector/target cells) was used. CA/EGFP LV was generated in the puc2CL11Egw.sh7FA-A7 with 2 sites BMF model in Fanca–/– mice. Mice with a targeted disruption in the of interference over the FANCA messenger. The shSCR/EGFP LV was Fanca gene (49) were a gift of H.J. van de Vrugt (Free University Med- used as an interference control and contained a scrambled sequence ical Center, Amsterdam, Netherlands) and were bred and maintained and the EGFP marker gene. The functionality of the shFANCA/EGFP at our facility as previously described (50). Six- to 8-month-old female LV has been previously demonstrated (47). HD CB CD34+ cells were FVB-129 WT and Fanca–/– mice were maintained under high-standard stimulated for 2 days in IMDM plus GlutaMAX supplemented with 20% conditions and routinely screened for pathogens. The induction of FBS and the hematopoietic growth factors stem cell factor (SCF) (30 ng/ BMF in Fanca–/– mice was performed by i.p. administration of 2 doses mL), thrombopoietin (TPO) (10 ng/mL), and FMS-like tyrosine kinase of 0.3 mg/kg MMC (MilliporeSigma), spaced 7 days apart. To evaluate 3 ligand (FLT3-L) (10 ng/mL) (Miltenyi Biotec) in a humidified atmo- the relevance of NKG2D–NKG2D-L interactions to the BMF in these sphere at 37°C and 5% CO2. Cells were transduced with the shSCR/EG- animals, mice were treated (i.p.) every 2–3 days with an InVivoMAb FP or shFANCA/EGFP LVs using 2 infection cycles of 24 hours in plates anti–mouse NKG2D antibody (CD314, clone CX5) or an isotype con- precoated with Retronectin (Takara). Transduced cells were then incu- trol InVivoMAb rat IgG1 (anti-HRP) antibody at 10 mg/kg (both from bated for 4–12 days in culture under conditions used for prestimulation. Bio X Cell). These antibodies were diluted in InVivoPure pH 7 dilution Levels of NKG2D-Ls in cells transduced with the shFANCA/EGFP LV buffer (Bio X Cell) for i.p. administration. The experimental protocol were compared with shSCR/EGFP-transduced cells showing similar used to evaluate the effect of NKG2D–NKG2D-L interactions on the levels of EGFP expression, and also with EGFP– cells. BMF in Fanca–/– mice is described in Figure 7A. CHEK-1 Western blotting. To analyze the inhibitory activity of To evaluate the course of the BMF, PB samples (100 μL) were peri- SB-2118078 on CHEK-1 activation, primary fibroblasts from HDs odically obtained from the tail vein and anticoagulated with EDTA and patients with FA were cultured with the inhibitor for 24 hours (1 (Microvette CB300, Sarstedt). PB parameters were determined in a μM) and then treated with trypsin-EDTA (1%) and washed in PBS. hematological analyzer (Sysmex XN-1000). At the end of the treatment, To obtain cytoplasmic and nuclear protein in denaturing conditions, the mice were sacrificed, and femoral BM cells were resuspended in cells were lysed with RIPA lysis buffer (Thermo Fisher Scientific) IMDM medium (Gibco Laboratories, Thermo Fisher Scientific) supple- containing 25 mM Tris-HCL, pH 7.6, 150 mM NaCl, 1% NP-40, 1% mented with 10% FBS (Sigma Chemical, MilliporeSigma). A mixture of sodium deoxycholate, and 0.1% SDS with protease and phosphatase anti–mRae-1-PE pan-specific (FAB17582P), anti–mH-60-PE (FAB1155P), inhibitors. Lysates were immunoblotted using a rabbit mAb against and anti–mMULT-1-PE (FAB2588P) antibodies (all from R&D Systems) phosphorylated CHEK-1 (Ser296) (Abcam, catalog ab79758), a rabbit was used to evaluate the expression of NKG2D-Ls in mice, as described anti-vinculin antibody (Abcam, catalog ab129002), and a goat poly- above for the analyses of human cells. Analyses of mouse hematopoiet- clonal anti-rabbit antibody (HRP) (Abcam, ab6721), and revealed with ic progenitor cells (Lin–c-Kit+) and more primitive Lin–Sca1+cKit+ (LSK) Super Signal West Pico PLUS (Thermo Fisher Scientific). precursors were performed using the following mAbs for lineage-posi- Incubation of the CHEK-1 inhibitor with FANCA-like CD34+ cells. To tive staining: anti–Gr-1 (clone RB6-8C5, BioLegend), anti-B220 (clone analyze the role of the DNA damage sensor CHEK-1 in the expression RA3-6B2, BioLegend), anti-CD3 (clone 145-2C11, BD Pharmingen), of NKG2D-Ls in FA-like HSPCs, CD34+ cells were transduced with anti-CD11b (clone M1/70, eBioscience), and anti-Ter119 (clone TER119, shFANCA/EGFP (see above), and 7 days later, cells were treated for BD Pharmingen), all with FITC. For positive staining of HSPCs, anti– 1–2 additional days with the CHEK-1 inhibitor SB-218078 (Calbio- Sca-1–APC–Cy7 (clone D7, BioLegend) and anti–c-Kit–PE–Cy5 (clone chem) dissolved in DMSO at a concentration of 1 μM (48). NKG2D-Ls 2B8, eBioscience) were selected. Immune cells were analyzed using the levels were then analyzed in CD34+EGFP+ cells and compared with following mAbs: anti-NK1.1 (clone PK136), anti-CD4 (clone H129.19), control cells only treated with DMSO. and anti-CD8 (clone 53–6.7), all with FITC and all from BD Pharmingen. In vitro NKG2D-blocking assays. This blocking assay was based on To evaluate the immune cell activation status, these cells were stained methods previously reported (15). HES-fractionated BM cells from with anti–CD69-PE (clone H1.2F3) from BD Pharmingen. HDs and patients with FA were divided into 2 equal aliquots (range: Statistics. Normal distribution was analyzed by the Shapiro- 0.5 × 106 to 1 × 106 cells) and resuspended in 200 μL IMDM supple- Wilks test. To determine statistically significant differences between mented with 20% FBS and loaded into V-bottomed 96 MicroWell independent continuous variables, a 2-tailed, paired or unpaired Stu- plates (Nunc). One of the aliquots was incubated with mouse IgG1 dent’s t test with Welch’s correction was applied when an equal SD was (25 μg/mL) as an isotype control (clone 11711), and the other one not assumed for normal distribution, and a 2-tailed, nonparametric Mann was incubated with the anti-NKG2D antibody (25 μg/mL) to block Whitney U test was used for non-normal distribution. For multiple com- the NKG2D-activating receptor (clone 149810), both from R&D Sys- parisons, the Kruskal-Wallis test was followed by Dunn’s multiple-com- tems. Cells were incubated for 10 minutes at RT and resuspended parison test. Spearman’s rank correlation coefficient was used to analyze after 5 minutes. Subsequently, cells were gently centrifuged at 60g associations between 2 continuous variables. Statistical analyses were for 3 minutes to facilitate cell interaction at the well bottom and were performed with GraphPad Prism, version 7.00.159 (GraphPad Software), then further incubated at 37°C in 5% CO2 for 4 hours. Cells were then and P values of less 0.05 were considered statistically significant. 14 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 The Journal of Clinical Investigation   RESEARCH ARTICLE Study approval. All experimental procedures were carried out Acknowledgments according to Spanish and European regulations. The experimen- The authors would like to thank Aurora de la Cal for her assis- tal mice procedures were approved by the committee at the CIE- tance in coordinating the delivery of patients’ samples to physi- MAT facility in Madrid, Spain (PROEX 156.5/21, regulation no. cians from the Spanish Network for Fanconi Anemia; Miguel A. 28079-21 A). The study involving patients with FA was approved by Martín for careful assistance with the mouse experiments; and the Centro de Investigación Biomédica en Red de Enfermedades Mercedes López-Santalla for help with statistical analyses. This Raras (CIBERER) (CIBERER-16-12-2010 and CIBER 17-6-2015). work was supported by the “Ministerio de Ciencia e Innovación Patients with FA and HDs provided written informed consent pri- y Competitividad y Fondo Europeo de Desarrollo Regional (FED- or to participation in this study. Studies conducted with samples ER)” (SAF2015-68073-R, SAF2015-64152-R, and RTI2018- from HDs and patients with FA were approved by the IRB of the Hos- 097125-B-I00); Next Generation EU; Plan de Recuperación pital del Niño Jesús (Madrid, Spain) and according to the Declaration Transformación y Resilencia (Instituto de Salud Carlos III; TER- of Helsinki. Patient and HD identities were encoded to protect con- AV) (RD12/0019/0023); Programs of the European Commission fidentiality. In all instances, written informed consent was obtained (HEALTHF5-2012-305421 and EUROFANCOLEN); the “Minis- from the study participants. terio de Sanidad, Servicios Sociales e Igualdad” (EC11/060 and EC11/550 “Comunidad de Madrid” (AvanCell, B2017/BMD- Author contributions 3692); and the ICREA-Academia program. JAC conceived the study. AV, RSD, PV, AL, SN, OA, RP, JCS, JM, and J Surrallés performed the experiments. HH, provided key vec- Address correspondence to: Juan Bueren or Jose A. Casado, Bio- tors for the study. CDH and J Sevilla provided critical samples and medical Innovation Unit; Division of Hematopoietic Innovative studied the patients. JAB directed the study and, with JAC, ana- Therapies, Av Complutense 40, CIEMAT, 28040 Madrid, Spain. lyzed the data and wrote the manuscript. All authors contributed Phone: 34.91.804.2525; Email: juan.bueren@ciemat.es (JB); jose. to the discussion and writing of the manuscript. casado@ciemat.es (JAC). 1. Kutler DI, et al. A 20-year perspective on the topoietic stem cell defects and bone marrow evolutionary biology of NKG2D ligands. Immunol International Fanconi Anemia Registry (IFAR). failure in fanconi anemia. Cell Stem Cell. Rev. 2015;267(1):88–116. Blood. 2003;101(4):1249–1256. 2016;18(5):668–681. 25. Gasser S, Raulet DH. The DNA damage response 2. Ebens CL, et al. Hematopoietic cell transplan- 14. Sumpter R Jr., et al. Fanconi anemia proteins arouses the immune system. Cancer Res. tation in Fanconi anemia: current evidence, function in mitophagy and immunity. Cell. 2006;66(8):3959–3962. challenges and recommendations. Expert Rev 2016;165(4):867–881. 26. Iannello A, Raulet DH. Immune surveillance of Hematol. 2017;10(1):81–97. 15. Hanaoka N, et al. NKG2D-mediated immunity unhealthy cells by natural killer cells. Cold Spring 3. Rio P, et al. Successful engraftment of gene-cor- underlying paroxysmal nocturnal haemoglobin- Harb Symp Quant Biol. 2013;78:249–257. rected hematopoietic stem cells in non-condi- uria and related bone marrow failure syndromes. 27. Cerboni C et al. The DNA damage response: a tioned patients with Fanconi anemia. Nat Med. Br J Haematol. 2009;146(5):538–545. common pathway in the regulation of NKG2D 2019;25(9):1396–1401. 16. Mavroudi I, Papadaki HA. Genetic associations and DNAM-1 ligand expression in normal, infect- 4. Ceccaldi R, et al. The Fanconi anaemia pathway: in acquired immune-mediated bone marrow fail- ed, and cancer cells. Front Immunol. 2014;4:508. new players and new functions. Nat Rev Mol Cell ure syndromes: insights in aplastic anemia and 28. Zingoni A, et al. NKG2D and its ligands: “one for Biol. 2016;17(6):337–349. chronic idiopathic neutropenia. Clin Dev Immu- all, all for one”. Front Immunol. 2018;9:476. 5. Niraj J, et al. The fanconi anemia pathway in can- nol. 2012;2012:123789. 29. Paczulla AM, et al. Absence of NKG2D ligands cer. Annu Rev Cancer Biol. 2019;3:457–478. 17. Boudreau JE, Hsu KC. Natural killer cell edu- defines leukaemia stem cells and mediates their 6. Bagby G. Recent advances in understanding hema- cation in human health and disease. Curr Opin immune evasion. Nature. 2019;572(7768):254–259. topoiesis in Fanconi Anemia. F1000Res. 2018;7:105. Immunol. 2018;50:102–111. 30. Gasser S, et al. The DNA damage pathway regu- 7. Joenje H, et al. Oxygen-dependence of chromo- 18. Lanier LL. NKG2D receptor and its ligands lates innate immune system ligands of the NKG2D somal aberrations in Fanconi’s anaemia. Nature. in host defense. Cancer Immunol Res. receptor. Nature. 2005;436(7054):1186–1190. 1981;290(5802):142–143. 2015;3(6):575–582. 31. Cerboni C, et al. Antigen-activated human T 8. Li Y, et al. Persistent response of Fanconi anemia 19. El-Gazzar A, et al. Immunobiology and con- lymphocytes express cell-surface NKG2D ligands haematopoietic stem and progenitor cells to oxi- flicting roles of the human NKG2D lymphocyte via an ATM/ATR-dependent mechanism and dative stress. Cell Cycle. 2017;16(12):1201–1209. receptor and its ligands in cancer. J Immunol. become susceptible to autologous NK- cell lysis. 9. Langevin F, et al. Fancd2 counteracts the toxic 2013;191(4):1509–1515. Blood. 2007;110(2):606–615. effects of naturally produced aldehydes in mice. 20. Diefenbach A, et al. Selective associations with 32. Cooper MA, et al. The biology of human Nature. 2011;475(7354):53–58. signaling proteins determine stimulatory versus natural killer-cell subsets. Trends Immunol. 10. Garaycoechea JI, et al. Genotoxic conse- costimulatory activity of NKG2D. Nat Immunol. 2001;22(11):633–640. quences of endogenous aldehydes on mouse 2002;3(12):1142–1149. 33. Melsen JE, et al. Human circulating and tis- haematopoietic stem cell function. Nature. 21. Gonzalez S, et al. NKG2D ligands: key targets sue-resident CD56(bright) natural killer cell 2012;489(7417):571–575. of the immune response. Trends Immunol. populations. Front Immunol. 2016;7:262. 11. Dufour C, et al. TNF-alpha and IFN-gamma are 2008;29(8):397–403. 34. Carreau M, et al. Bone marrow failure in the Fan- overexpressed in the bone marrow of Fanconi 22. Eagle RA, et al. Beyond stressed self: evidence for coni anemia group C mouse model after DNA anemia patients and TNF-alpha suppresses eryth- NKG2D ligand expression on healthy cells. Curr damage. Blood. 1998;91(8):2737–2744. ropoiesis in vitro. Blood. 2003;102(6):2053–2059. Immunol Rev. 2009;5(1):22–34. 35. Ceccaldi R, et al. Spontaneous abrogation of the 12. Briot D, et al. Aberrant activation of stress-response 23. Raulet DH, et al. Regulation of ligands for the G(2)DNA damage checkpoint has clinical bene- pathways leads to TNF-alpha oversecretion in Fan- NKG2D activating receptor. Annu Rev Immunol. fits but promotes leukemogenesis in Fanconi ane- coni anemia. Blood. 2008;111(4):1913–1923. 2013;31:413–441. mia patients. J Clin Invest. 2011;121(1):184–194. 13. Zhang H, et al. TGF-β inhibition rescues hema- 24. Carapito R, Bahram S. Genetics, genomics, and 36. Ceccaldi R, et al. Bone marrow failure in Fanconi J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842 15 RESEARCH ARTICLE The Journal of Clinical Investigation   anemia is triggered by an exacerbated p53/p21 T-cell trafficking: parallels and distinctions. 2005;12(5):976–984. DNA damage response that impairs hemato- Immunotherapy. 2011;3(10):1223–1233. 46. Hanenberg H, et al. Phenotypic correction of poietic stem and progenitor cells. Cell Stem Cell. 42. Scheinberg P, Young NS. How I treat acquired primary Fanconi anemia T cells with retrovi- 2012;11(1):36–49. aplastic anemia. Blood. 2012;120(6):1185–1196. ral vectors as a diagnostic tool. Exp Hematol. 37. Pontel LB, et al. Endogenous formaldehyde is a 43. Casado JA, et al. A comprehensive strategy 2002;30(5):410–420. hematopoietic stem cell genotoxin and metabolic for the subtyping of patients with Fanconi 47. Nalepa G, et al. Fanconi anemia signaling net- carcinogen. Mol Cell. 2015;60(1):177–188. anaemia: conclusions from the Spanish Fan- work regulates the spindle assembly checkpoint. 38. Stojanovic A, et al. The NKG2D/NKG2DL axis in coni Anemia Research Network. J Med Genet. J Clin Invest. 2013;123(9):3839–3847. the crosstalk between lymphoid and myeloid cells 2007;44(4):241–249. 48. Zhao B, et al. Structural basis for Chk1 inhibition by in health and disease. Front Immunol. 2018;9:827. 44. Bryant BJ, et al. Gravity sedimentation of UCN-01. J Biol Chem. 2002;277(48):46609–46615. 39. Ames E, et al. NK cells preferentially target tumor granulocytapheresis concentrates with 49. Cheng NC, et al. Mice with a targeted disruption cells with a cancer stem cell phenotype. J Immu- hydroxyethyl starch efficiently removes red of the Fanconi anemia homolog Fanca. Hum Mol nol. 2015;195(8):4010–4019. blood cells and retains neutrophils. Transfusion. Genet. 2000;9(12):1805–1811. 40. Parolini S, et al. The role of chemerin in the colo- 2010;50(6):1203–1209. 50. Fernández-García M, et al. Improved hematopoi- calization of NK and dendritic cell subsets into 45. Chandra S, et al. A rapid method for retrovi- etic gene therapy in a mouse model of fanconi inflamed tissues. Blood. 2007;109(9):3625–3632. rus-mediated identification of complementation anemia mediated by mesenchymal stromal cells. 41. Nolz JC, et al. Naive, effector and memory CD8 groups in Fanconi anemia patients. Mol Ther. Hum Gene Ther. 2018;29(3):327–336. 16 J Clin Invest. 2022;132(15):e142842 https://doi.org/10.1172/JCI142842