Csa Can Induce DNA Double-Strand Breaks: Implications for BMT Regimens Particularly for Individuals with Defective DNA Repair

ORIGINAL ARTICLE CsA can induce DNA double-strand breaks: implications for BMT regimens particularly for individuals with defective DNA repair

M O’Driscoll and PA Jeggo

Genome Damage and Stability Centre, University of Sussex, Brighton, East Sussex, UK

Several human disorders mutated in core components of myelo-ablative haematopoietic stem cell transplantation the major DNA double-strand break (DSB) repair path- (NHSCT) using reduced doses of DNA damaging agents way, non-homologous end joining (NHEJ), have been has improved the success rate for transplantation in this described. Cell lines from these patients are characterized context.1,3 Several human disorders have now been by sensitivity to DSB-inducing agents. DNA ligase IV described that are defective in non-homologous end-joining syndrome (LIG4) patients specifically, for unknown (NHEJ), the principal pathway by which human cells repair reasons, respond particularly badly following treatment DNA double-strand breaks (DSBs).4,5 These include DNA for malignancy or BMT. We report the first systematic ligase IV syndrome (LIG4), which is caused by hypo- evaluation of the response of LIG4 syndrome to morphic mutations in DNA ligase IV, Artemis-dependent compounds routinely employed for BMT conditioning. SCID (ART-SCID), caused by mutations in the Artemis We found human pre-B lymphocytes, a key target endonuclease and XRCC4-like factor (XLF)/Cernunnos- population for BMT conditioning, when deficient for dependent SCID, caused by mutations in an XLF/ DNA ligase IV, unexpectedly exhibit significant sensitiv- Cernunnos.5–8 Increased cellular and clinical radiosensitivity ity to CsA the principal prophylaxis for GVHD. is a feature of these conditions due to their failure to repair Furthermore, we found that CsA treatment alone or in ionizing radiation-induced DSBs.6 Because of the central combination with BU and fludarabine resulted in increased role played by NHEJ in V(D)J recombination all of these levels of DSBs specifically in LIG4 syndrome cells disorders initially present with moderate-to-severe com- compared to wild-type or Artemis-deficient cells. Our bined immunodeficiency necessitating clinical intervention study shows that CsA can induce DSBs and that LIG4 often culminating in BMT.9 But, anecdotal (unpublished syndrome patient’s fail to adequately repair this damage. data) and evidence from other studies suggest that LIG4 These DSBs likely arise as a consequence of DNA syndrome patients are particularly poorly following BMT replication in the presence of CsA. This work has (Table 1). Indeed, there is only one detailed report of a implications for BMT and GVHD management in general successful BMT in LIG4 syndrome using NHSCT with a and specifically for LIG4 syndrome. modified conditioning regimen.12 This is in marked contrast Bone Marrow Transplantation (2008) 41, 983–989; to that of ART-SCID where NHSCT has been used doi:10.1038/bmt.2008.18; published online 18 February 2008 successfully in these patients even before the underlying Keywords: CsA; DNA double-strand breaks; LIG4 causative genetic defect was identified.9,13,14 The reason for syndrome the difference in response in these two patient groups that display similar levels of radiosensitivity is unclear. To date, there has not been a report of BMT in XLF/Cernunnos- dependent SCID. The agents commonly employed as part of the con- Introduction ditioning regimen prior to NHSCT include BU (Myerlan), a bi-functional DNA crosslinking agent, fludarabine BMT remains one of the only curative options available to (FluD) (Fludara), a nucleoside analogue and MTX treat progressive anaemia, combined immunodeficiency (Amethopterin), a dihydrofolate reductase inhibitor. All of and/or lymphoma development in the context of DNA these compounds can induce DSBs either directly or damage response-defective disorders, such as Fanconi indirectly. CsA is a widely used prophylaxis for GVHD, anaemia and Nijmegen breakage syndrome.1,2 Non- which represents a significant post transplantation compli- cation.15,16 CsA binds to the cyclophilin class of proteins that can then function as an inhibitor of calcineurin, a Correspondence: Dr M O’Driscoll, Genome Damage and Stability serine-threonine phosphatase (also called PP2B). CsA’s Centre, University of Sussex, Science Park Road, Falmer, Brighton, utility as an immunosuppressive agent is thought to derive East Sussex, BN1 9RQ, UK. E-mail: [email protected] from its ability to prevent nuclear localization of transcrip- Received 1 November 2007; revised 9 January 2008; accepted 10 January tion factor NFAT (nuclear transcription factor in activated 2008; published online 18 February 2008 T cells).17 NFAT is involved in the transcription of various CsA, DSBs and LIG4 syndrome M O’Driscoll and PA Jeggo 984 Table 1 A summary of the outcome of clinical management of BMT on ART-SCID and LIG4 syndrome patients

Publication Details

ART-SCID O’Marcaigh et al.9 420 Patients successfully transplanted using NHSCT. Treated ﬁve patients with CsA only 3 mg/kg per day. For GVHD prophylaxis CsA (3 mg/kg per day) and MTX, initially at 0.5 mg/kg, and the remaining three at 0.33 mg/kg given on days 1, 3, 6 and 11 post transplant

LIG4 Buck et al.5 Patient no. 1: BU, EDX, ATG Â 2. The donor was her mother (3/6 identical). Patient developed moderate VOD and eventually an EBV-induced lymphoproliferation and died Patient no. 2: Transplanted at 2.5 months of age receiving 5 days of ATG 5 mg/kg per day from day À14 to À10, then BU 2 mg/kg for 12 h p.o. during 4 days, followed by EDX 50 mg/kg per day from day À5 to day À2. Patient received CD34-positive selected marrow from her mother (3/6 identical). Patient developed severe VOD and died of complications

Enders et al.10 Patient no. 1: Transplanted at 4 months of age with Thiotepa 3 Â 5 mg/kg (day À8 to day À6), Fludarabine 4 Â 1.43 mg/kg (day À5 to day À2), ATG 3 Â 20 mg/kg (day À3 to day À1), MUD, 1.87 Â 107 CD 34 per kg. For GVHD prophylaxis CsA from day À1 (3 mg/kg per day) and MTX (10 mg/kg) day +1, 3, 6, 11. Patient developed hemolytic-uremia (anemia and renal failure) 3 months post transplantation assumed to have been as an overresponse to the GVHD treatment Patient no. 2: Died following treatment for EBV-NHL using the NHL-BFM-95 protocol with modiﬁcations (1 g instead of 5 g MTX, no ifosphamid) with additional CSF therapy due to blasts in the CSF (MTX and prednisone)

van der Burg et al.11 This patient died of possible VOD following treatment with BU (50 mg per day) and cyclophosphamide (500 mg per day)

Gruhn et al.12 This patient was diagnosed as LIG4 prior to transplantation hence a NHSCT protocol was applied.6 This patient was transplanted at 10.5 years of age for pancytopenia, Fludara 4 Â 30 mg/m2 per day, EDX 4 Â 10 mg/kg per day, ATG 4 Â 15 mg/m2 per day donor: HLA-identical brother. For GVHD prophylaxis CsA (3 mg/kg per day) from day À1 to d+64, G-CSF from day +5

Abbreviations: EBV-NHL ¼ Epstein–Barr virus-positive non-Hodgkin lymphoma; EDX ¼ endoxan (cyclophosphamide); MUD ¼ matched unrelated donor; NHSCT ¼ non-myeloablative haematopoietic SCT; VOD ¼ veno-occlusive disease.

cytokines (for example IL-2) that are required to activate described elsewhere.6,19 The LIG4 syndrome fibroblast was T cells. Inhibition of NFAT nuclear localization by CsA derived from a 9-year-old patient with three homozygous results in profound systemic immunosuppression. alterations in LigIVÀ/À (8C4T(A3V), 26C4T(T9I) and Because of the poor clinical outcome of LIG4 syndrome 833G4A(R278H)).6 The ART-SCID fibroblast (CJ179) patients to BMT we set out to investigate whether the was derived from an immuno-deficient child and fails to failure to repair DSBs induced by the agents used during express a detectable Artemis transcript due to a genomic NHSCT could have a disproportionately adverse impact in deletion.19 the context of LigIVÀ/À deficiency. Using the methylthiazolyldiphenyl-tetrazolium bromide (MTT) assay, we eval- Drugs uated the sensitivity of human pre-B lymphocytes either All drugs/chemicals were obtained from Sigma-Aldrich UK proficient or engineered by gene targeting to be deficient for Ltd (Poole, UK) and made up fresh prior to each survival. À/À LigIV to the components employed clinically as part of BU was dissolved in DMSO and made up to 5 mM in PBS the conditioning regimen prior to BMT and for GVHD (p2% v/v DMSO). FluD and CsA were dissolved in prophylaxis. These included BU, FluD, MTX and CsA. DMSO. Methyl methane sulphonate (MMS) was diluted in Furthermore, using an indirect immunofluorescence assay complete medium. based on 53BP1 foci formation as a sensitive marker for DSBs, we evaluated the impact of BMT conditioning regimens on DSB formation and repair in NHEJ-defective Antibodies human cells. Anti-53BP1 (BL181) antibody was obtained from Uni- versal Biologicals (Cambridge, UK). Anti-bromodeoxyuridine (BrdU; Bu20A) antibody was obtained from Autogen BioClear (Wiltshire, UK). Materials and methods Immunofluorescence-based DNA DSB repair assay Cell lines Immunofluorescence-based detection of 53BP1 foci forma- Wild-type (WT; Nalm 6) and LigIVÀ/À (N114-P2) pre-B tion using an anti-53BP1-specific antibody was used as an human lymphocyte cells have been described previously.18 indirect highly sensitive assay for monitoring DNA DSB These cells were cultured in RPMI-1640 supplemented with formation and repair, as previously characterized.19 53BP1 L-glutamine and 15% FCS. Primary human skin fibroblasts is one of many proteins recruited to a DSB. Following were grown in MEM supplemented with L-glutamine and treatment of primary fibroblasts with various agents, single 15% FCS. The WT (1BR.3), LIG4 syndrome (LIG4; DSBs are visualized indirectly as discrete microscopically 411BR) and Artemis-SCID (ART-SCID; CJ179) have been detectable 53BP1 foci. Cells were also treated with 50 mM

Bone Marrow Transplantation CsA, DSBs and LIG4 syndrome M O’Driscoll and PA Jeggo 985 BrdU to label S-phase cells that were specifically visualized for conditioning regimens. We treated WT and Lig IVÀ/À following immunofluorescence using an anti-BrdU anti- human pre-B lymphocytes with BU, FluD and MTX either body. Standard immunofluorescence procedures for alone (Figure 1a–c) or in various combinations (Figure 1d) primary skin fibroblast have been described in detail and determined survival after 5 days using MTT analysis. elsewhere.19 Images were captured using a Zeiss-Axioplan We failed to find any significant sensitivity in the absence of microscope using Simple-PCI software. Lig IVÀ/À to any of these agents individually (Figure 1a–c). Furthermore, no additional significant sensitivity was observed in LigIVÀ/À pre-B lymphocytes when they were MTT assay treated with all of these agents simultaneously (Figure 1d). For MTT survival analysis cells were treated in RPMI-1640 This data suggest that complete loss of LigIVÀ/À does not without phenol red (Fischer Scientific UK Ltd, Lough- selectively significantly sensitize human pre-B lymphocytes borough, Leicestershire, UK) in 24-well plates and grown for to BU, FluD or MTX. Furthermore, when used in 5 days. The drug(s) was not removed. MTT dissolved in combination, these compounds do not act to synergistically RPMI-1640 without phenol red was added to each well sensitize human pre-B lymphocytes deficient for LigIVÀ/À. (0.5 mg/ml) and incubated for 3–6 h to allow the insoluble CsA is currently a cornerstone for the treatment and/or formazan crystals to form. Acidic iso-propanol (40 mM HCl prevention of GVHD.16 Unexpectedly, we found that Lig in absolute iso-propanol) was added to each well to dissolve IVÀ/À human pre-B lymphocytes were selectively sensitive the formazan crystals. Following removal of cell debris by to killing by CsA (Figure 2a). Similarly, when CsA was centrifugation, absorbance of the converted dye was used in combination with BU and FluD, we found that the measured at 570 nm with background subtraction at 650 nm. LigIVÀ/À pre-B lymphocytes were more sensitive than their WT counterparts (Figure 2b). Since LigIVÀ/À functions as a core component in one of the principal DNA DSB repair Results pathways, NHEJ, we investigated the possibility that this selective sensitivity of LigIVÀ/À cells towards CsA could be We specifically chose human pre-B lymphocytes for the consequence of CsA-induced DSB formation. It has not analysis as these represent an important target population been previously reported that CsA can, either directly or

100 100 WT WT 80 80 LigIV –/– LigIV –/–

60 60

40 40 Survival (%) Survival (%) 20 20

0 0 0 100 200 300 400 500 0 2 4 6 8 10 12 [Busulphan] µM [Fludarabine] µM

100 120 WT LigIV –/– WT 100 80 LigIV –/– 80 60 60

40 Survival (%) Survival (%) 20 20 0

0 BU UNT FluD MTX 0 0.05 0.1 0.15 0.2 0.25 0.3 Bus+Flud Bus+MTXFluD+MTX [Methotrexate] µM Bus+FluD+MTX Figure 1 DNA ligase IVÀ/À (LigIVÀ/À) human pre-B lymphocytes are not selectively sensitive to killing by BU, FluD or MTX, alone or in combination. Wild-type (WT) and LigIVÀ/À human pre-B lymphocytes (LigIVÀ/À) were chronically exposed to increasing concentrations of BU (a), FluD (b) and MTX (c) in culture for 5 days. Survival was determined using the methylthiazolyldiphenyl-tetrazolium bromide (MTT) assay. Drugs were not removed during the / course of the experiment. (d) WT and LigIVÀ À human pre-B lymphocytes were treated simultaneously with different combinations of BU (5 mM), FluD (0.05 mM) and MTX (0.025 mM) to investigate the potential for these agents to kill these cells synergistically. Cells were incubated for 5 days prior to analysis by MTT. Drugs were not removed.

Bone Marrow Transplantation CsA, DSBs and LIG4 syndrome M O’Driscoll and PA Jeggo 986 100 compared to WT and ART-SCID (Figure 3a and b). 53BP1 WT foci were not observed in cells that did not stain for BrdU, 80 LigIV –/– under these conditions. Staining for S-phase cells using anti-BrdU showed that all of the cell lines had comparable 60 levels of S-phase cells (20–23%) and were therefore growing at similar rates. This indicates that CsA treatment results in DSB formation in cells that have traversed S-phase 40 and that these breaks persist in LIG4 syndrome cells Survival (%) specifically. 20 A likely mechanism for DSB formation during S phase is the collision of replication forks with single-stranded DNA 0 breaks (SSBs). SSBs occur spontaneously at high levels in 0 24681012 mammalian cells but are usually repaired rapidly by the [CSA] µM base excision repair pathway (BER).24–27 CsA has been proposed to inhibit DNA damage-induced overexpression 20 of DNA polymerase b, which plays an important role in BER.26,28 To examine whether CsA might enhance the 15 persistence or formation of DSBs following replication, we examined the impact of CsA on DSBs arising following 10 treatment with MMS, an agent that can induce SSBs in S-phase (Figure 3c). Following exposure to 1 mM MMS 5 for 1 h and incubation for a further 24 h, a small increase in DSBs (5–8 53BP1 foci) was observed in WT, LIG4 Relative survival (%) 0 syndrome and ART-SCID primary fibroblasts that had WT LigIV –/– traversed S-phase (BrdU þ ive). Non-BrdU-labelled cells showed no increase in DSB formation suggesting that DSB À/À À/À Figure 2 DNA ligase IV (LigIV ) pre-B lymphocytes are selectively formation is replication dependent. This is consistent with sensitive to CsA. (a) The survival of wild-type (WT) and LigIVÀ/À human pre-B lymphocytes (LigIVÀ/À) to increasing concentrations of CsA was the notion that DSBs can arise following replication of determined by methylthiazolyldiphenyl-tetrazolium bromide (MTT) assay SSBs but that under normal conditions they are rapidly 5 days post-treatment. CsA was not removed. (b) WT and LigIVÀ/À human repaired. Strikingly, when CsA (5 mM) was included with pre-B lymphocytes were treated simultaneously with BU (5 mM), FluD MMS, a significant increase in DSBs was observed, (1 mM) and CsA (2 mM) and survival determined after 5 days culture using specifically in LIG4 syndrome fibroblasts (23–30 53BP1 the MTT assay. Drugs were not removed. foci) compared to WT or ART-SCID (5–10 53BP1 foci). This suggests that CsA-induced DSBs can form, likely indirectly, from SSBs during S-phase. Furthermore, these indirectly, induce DSBs. Following DSB formation in DSBs specifically persist in LIG4 syndrome cells compared mammalian cells, one of the earliest detectable responses to ART-SCID (Figure 3c). is the phosphorylation (ATM and DNA-protein kinase Finally, we examined whether the increased sensitivity of dependent) of the histone H2A variant H2AX (termed g- LigIVÀ/À cells to combined treatment with CsA, BU and H2AX). It has been shown that DSB formation and repair FluD, conditions which mimic NHSCT conditioning can be very sensitively monitored in cells by indirect and GVHD prophylaxis, was a consequence of increased immunofluorescence microscopy using antibodies specific DSB formation (Figure 4). Using WT and LIG4 syndrome for g-H2AX.19 It has been shown that each discrete primary fibroblasts, following treatment with modest microscopically detectable g-H2AX focus likely represents concentrations of BU (5 mM) together with FluD (1 mM), a single DSB.20–22 To monitor DSB formation we used no significant difference in DSBs was observed between primary skin fibroblasts that are more amenable to this these cell lines (Figure 4; BU þ FluD). This was distinct to type of microscopic analysis compared to lymphocytes. treatment with CsA (5 mM) alone where an increased level of 53BP1 is one of many proteins recruited very rapidly to a DSBs were observed in LIG4 syndrome cells derived from DSB. Microscopically detectable 53BP1 foci are formed, S-phase specifically (Figure 4 CsA and Figure 3b). Strik- co-localize and are repaired with the same kinetics as those ingly, when these cell lines were treated for 24 h with CsA 19,23 of g-H2AX. For practical reasons (to allow co-staining (5 mM) simultaneously with BU (5 mM) and FluD (1 mM), a with anti-BrdU) we utilized 53BP1 foci formation as a significant increased level of DSBs were now observed surrogate marker for DSBs instead of monitoring g-H2AX specifically in the LIG4 syndrome cells compared to WT foci formation (Figure 3a). We used exponentially growing (Figure 4; BU þ FluD þ CsA). This data suggest that CsA primary fibroblasts from a clinically unaffected (WT), a can induce DSBs alone but particularly in combination radiosensitive DSB repair-defective LIG4 patient and with BU and FluD and that these DSBs remain at an Artemis-defective SCID patient (ART-SCID) that we elevated level in LIG4 syndrome-derived patient cells. previously characterized.6,19 Interestingly, we found an Collectively our data identify CsA as an important increased amount of DSBs (53BP1 foci) in cells that had source of DSBs especially when used in combination traversed S-phase (BrdU positive ( þ ive)) following a 24 h with BU and FluD, similarly to conditions used for treatment with CsA specifically in the LIG4 syndrome cells NHSCT conditioning and GVHD prophylaxis. Furthermore,

Bone Marrow Transplantation CsA, DSBs and LIG4 syndrome M O’Driscoll and PA Jeggo 987 12 WT LIG4 UNT 10 CSA

DAPI 8

4 53BP1 2 53BP1 Foci/BrdU +ive cell 0 WT LIG4 ART-SCID

35 WT 30 LIG4 ART-SCID 25 20 15 10 5 53BP1 Foci/BrdU +ive cell 0 UNT CSA MMS MMS+CSA Figure 3 CsA induces more double-strand breaks (DSBs) in DNA ligase IV syndrome (LIG4) cells compared to Artemis-dependent severe combined immunodeficiency (ART-SCID) and these DSBs likely arise from single-strand breaks during DNA replication. (a) Wild-type (WT) and LIG4 primary skin fibroblasts were treated with CsA (5 mM) for 24 h. DSB formation was monitored indirectly by examining the formation of 53BP1 foci at the site of DSBs by indirect immunofluorescence using an antibody against 53BP1. DSBs were visualized as discrete microscopically detectable 53BP1 foci. Cells were counterstained with a DNA-specific stain (4-6-diamidino-2-phenylindole, DAPI) to identify the nucleus. (b) WT, LIG4 and Artemis-defective SCID (ART- SCID) primary skin fibroblasts were untreated (UNT) or treated with CsA (5 mM). 53BP1 foci formation in bromodeoxyuridine-positive cells (BrdU þ ive) was determined 24 h post-treatment. All cell lines were also initially treated with 50 mM BrdU to label S-phase cells that were specifically identified as the BrdU þ ive staining by immunofluorescence. CsA and BrdU were not removed during the course of the experiment. (c) WT, LIG4 and ART-SCID primary skin fibroblasts were UNT treated with methylmethane sulphonate (MMS) and/or CsA (5 mM) and examined for 53BP1 foci formation in BrdU þ ive cells 24 h post-treatment. All cell lines were also initially treated with 50 mM BrdU. MMS treatment (1 mM) was for 1 h. Neither CsA nor BrdU was removed during the course of the experiment.

14 Discussion

12 WT NHSCT using reduced doses of DNA damaging agents has 10 LIG4 significantly improved the outcome of BMT in individuals compromised for DNA repair.1,14 But, LIG4 syndrome 8 patients, particularly compared to ART-SCID, generally 5,9–11,13,14 6 have a poorer outcome. Multiple factors could be responsible for this, especially the general health of the 4 patients prior to conditioning. Nevertheless, since com- 2 pounds that can form DSBs are used during NHSCT, we 53BP1 foci/BrdU +ive cells sought to determine whether LigIVÀ/À deficiency could 0 adversely hypersensitize individuals to such treatments. UNT Bus+FluD CSA Bus+Flud+CSA Here, we report the first detailed evaluation of the response of LigIVÀ/À-defective cells to the compounds used for BMT Figure 4 Co-treatment of DNA ligase IV syndrome (LIG4) cells with À/À CsA, BU and fludarabine (FluD) results in increased levels of double- and GVHD prophylaxis. We found that human LigIV strand breaks (DSBs) compared to Wild-type (WT) cells. There was no pre-B lymphocytes are not significantly selectively sensitive evidence of DSB formation in non-bromodeoxyuridine (BrdU)-labelled to BU, FluD or MTX alone or when treated with these cells. 53BP1 foci were monitored in cells that had traversed S-phase 24 h agents simultaneously. Unexpectedly, we found that following treatment with different combinations of BU (5 mM), FluD (1 mM) LigIVÀ/À pre-B human lymphocytes were selectively and CsA (5 mM) in WT and LIG4 primary skin fibroblasts. All cell lines were initially treated with 50 mM BrdU as above and drugs were not sensitive to CsA, probably the most widely used prophy- removed. BrdU þ ive ¼ bromodeoxyuridine-positive (S-phase) cells. laxis for GVHD. Furthermore, when CsA was used in combination with BU and FluD, these cells exhibited LigIVÀ/À pre-B lymphocytes and LIG4 syndrome primary significant hypersensitivity. Crucially, we found that CsA fibroblasts are particularly sensitive to CsA-induced DSB treatment resulted in DSB formation even in non-haema- formation. topoetic human cells (primary skin fibroblasts). We suggest

Bone Marrow Transplantation CsA, DSBs and LIG4 syndrome M O’Driscoll and PA Jeggo 988 that these DSBs arise likely indirectly from SSBs occurring induces persistent DSBs this could compromise genomic during DNA replication. Importantly, we showed that stability not just in the context of LigIVÀ/À deficiency. these CsA-induced DSBs are formed at increased levels in Whether this could play any role in BMT-derived LigIVÀ/À-deficient background following co-treatment with secondary cancers is unclear, although worthy of further BU and FluD. Finally we showed that these DSBs are investigation.32,43 repaired with slower kinetics in the absence of LigIVÀ/À compared to Artemis deficiency. Our data unexpectedly identified CsA as an agent Acknowledgements capable of causing DSB formation in human cells This work was carried out as part of a project grant funded by specifically following replication. Interestingly, CsA treat- the UK Leukaemia Research Fund to MO’D and PAJ. The ment, either alone or in combination with other BMT work was devised by MO’D and PAJ. MO’D carried out all of drugs, did not result in an increased level of DSBs in the experiments and wrote the article. Special thanks to David ART-SCID fibroblasts compared to those of LIG4 Grant (LRF) for supporting this project. MO’D’s laboratory is syndrome. The roles of Artemis and LigIVÀ/À in NHEJ funded by a Cancer Research UK Senior Research Fellowship are distinct.19,29 LIG4 syndrome cells show slow kinetics of and UK Medical Research Council. PAJ’s laboratory is repair of all DSBs, while ART-SCID cells rejoin the supported by the UK Medical Research Council, Human majority of DSBs with normal kinetics but are defective in Frontiers Science Program, UK Leukaemia Research Fund, a specific subset of DSBs. Increased levels of residual CsA- International Agency for Cancer Research, EU grant F16R-CT- induced DSBs seen in the LIG4 syndrome cells are likely a 2003-508842 (RiscRad) and 512113 (DNA Repair). Special thanks to all of the clinicians and scientists for discussing their result of their general inability to repair all types of DSBs LIG4 and ART-SCID BMT experiences, particularly Jean compared to ART-SCID cells. This provides a provocative Pierre de Villartay, Despina Moshous, Nico Hartwig, Morton potential explanation as to why ART-SCID individuals Cowan, Dik van Gent, Mirjam van der Burg, Stephan Ehl, routinely successfully undergo NHSCT compared to LIG4 Karen Cerosaletti and Sule Unal. syndrome. Our findings also add weight to the increasing evidence that NHEJ can function to repair replication- associated DSBs, which has previously been proposed to be Financial disclosure carried out by homologous recombination.30 CsA is used as prophylaxis for GVHD, but GVHD The authors declare no competing financial interests. treatment protocols vary significantly.16 A previous study among the European Group for Blood and Marrow References Transplantation (EBMT) found that CsA treatment protocols and target doses vary widely among European 1 Aschan J. Allogeneic haematopoietic stem cell transplantation: transplantation centres.31 While the average initial daily current status and future outlook. Br Med Bull 2006; 77–78: dose of CsA was 3 mg/kg (i.v.) the range was 1–20 mg/kg 23–36. (i.v.). Target CsA blood concentrations ranged from 250 to 2 Gennery A, Slatter M, Bhattacharya A, Jeggo P, Abinun M, Flood T et al. Bone marrow transplantation for Nijmegan 400 mg/l, which is equivalent to 0.19–0.3 mM CsA. While breakage syndrome. J Pediatr Hematol Oncol 2005; 27: 239. these low mM concentrations represent an approximately 3 Resnick IB, Shapira MY, Slavin S. Nonmyeloablative stem cell 10-fold difference to those used in our experiments transplantation and cell therapy for malignant and non- (2–5 mM), crucially, our cellular work utilized a short-term malignant diseases. Transpl Immunol 2005; 14: 207–219. exposure to a single dose of CsA. The cumulative effect on 4 O’Driscoll M, Gennery AR, Seidel J, Concannon P, Jeggo PA. DSB formation from repeated or chronic exposure to low An overview of three disorders associated with genetic mM concentrations of CsA could be physiologically instability: LIG4 syndrome, RS-SCID and ATR-Seckel relevant. Interestingly, the LIG4 syndrome patient that syndrome. DNA Repair 2004; 3: 1227–1235. successfully underwent NHSCT recently reported by 5 Buck D, Malivert L, de Chasseval R, Barraud A, Fondaneche M-C, Gruhn et al.,12 received a CsA dosage at the lower end of Sanal O et al. Cernunnos, a novel nonhomologous end-joining factor, is mutated in human immunodeficiency with micro- the European range (3 mg/kg per day).31 Our data suggest cephaly. Cell 2006; 124: 287–299. that caution should be exercised regarding CsA treatment 6 O’Driscoll M, Cerosaletti KM, Girard P-M, Dai Y, Stumm M, regimens specifically in the context of LIG4 syndrome. Kysela B et al. DNA ligase IV mutations identified in patients An important general implication of our findings that exhibiting development delay and immunodeficiency. Mol Cell CsA can form DSBs relates to malignancy. Long-term use 2001; 8: 1175–1185. of calcineurin inhibitors such as CsA in solid organ and 7 Moshous D, Callebaut I, de Chasseval R, Corneo B, BMT is associated with increased risk of cancer.32–34 While Cavazzana-Calvo M, Le Deist F et al. Artemis, a novel immunosuppression appears to be the principal predispos- DNA double-strand break repair/V(D)J recombination pro- ing factor, others have argued that compounds such as CsA tein, is mutated in human severe combined immune deficiency. can also inhibit the repair of DNA damage.33,35–37 A Cell 2001; 105: 177–186. precedent for the role of immunosuppressive agent-induced 8 Ahnesorg P, Smith P, Jackson SP. XLF interacts with the XRCC4–DNA ligase IV complex to promote DNA non- DNA damage likely impacting on cancer predisposition has homologous end-joining. Cell 2006; 124: 301–313. 38,39 recently been described regarding azathioprine. Un- 9 O’Marcaigh A, DeSantes K, Hu D, Pabst H, Horn B, Li L repaired DSBs can result in deletions and/or act as a et al. Bone marrow transplantation for T-B- severe combined platform for translocations with obvious implications for immunodeficiency disease in Athabascan-speaking native malignant transformation.40–42 If prolonged CsA treatment Americans. Bone Marrow Transplant 2001; 27: 703–709.

Bone Marrow Transplantation CsA, DSBs and LIG4 syndrome M O’Driscoll and PA Jeggo 989 10 Enders A, Fisch P, Schwarz K, Duffner U, Pannicke U, 28 Ahlers C, Kreideweiss S, Nordheim A, Ruhlmann A. Nikolopoulos E et al. A severe form of human combined Cyclosporin A inhibits Ca2+-mediated upregulation of the immunodeficiency due to mutations in DNA ligase IV. DNA repair enzyme DNA polymerase beta in human J Immunol 2006; 176: 5060–5068. peripheral blood mononuclear cells. Eur J Biochem 1999; 11 van der Burg M, van Veelen LR, Verkaik NS, Wiegant WW, 264: 952–959. Hartwig NG, Barendregt BH et al. A new type of radio- 29 Lobrich M, Jeggo PA. Harmonising the response to DSBs: a sensitive T-B-NK+ severe combined immunodeficiency new string in the ATM bow. DNA Repair 2005; 4: 749–759. caused by a LIG4 mutation. J Clin Invest 2006; 116: 137–145. 30 Takata M, Sasaki MS, Sonoda E, Morrison C, Hashimoto M, 12 Gruhn B, Seidel J, Zintl F, Varon R, Tonnies H, Neitzel H Utsumi H et al. Homologous recombination and non- et al. Successful bone marrow transplantation in a patient with homologous end-joining pathways of DNA double-strand DNA ligase IV deficiency and bone marrow failure. Orphanet J break repair have overlapping roles in the maintenance of Rare Dis 2007; 2:5. chromosomal integrity in vertebrate cells. EMBO J 1998; 17: 13 Kobayashi N, Agematsu K, Sugita K, Sako M, Nonoyama S, 5497–5508. Yachie A et al. Novel Artemis gene mutations of radiosensitive 31 Ruutu T, Niederwieser D, Gratwohl A, Apperley J. A survey severe combined immunodeficiency in Japanese families. of the prophylaxis and treatment of acute GVHD in Europe: a Hum Genet 2003; 112: 348–352. report of the European Group for Blood and Marrow 14 Buckley RH. Molecular defects in human severe combined Transplantation (EBMT). Bone Marrow Transplant 1997; 19: immunodeficiency and approaches to immune reconstitution. 759–764. Ann Rev Immunol 2004; 22: 625–655. 32 Deeg HJ, Socie G. Malignancies after hematopoietic stem cell 15 Deeg HJ, Leisenring W, StorbR, Nims J, Flowers MED, transplantation: many questions, some answers. Blood 1998; Witherspoon RP et al. Long-term outcome after marrow 91: 1833–1844. transplantation for severe aplastic anemia. Blood 1998; 91: 33 Buell JF, Gross TG, Woodle ES. Malignancy after transplant- 3637–3645. ation. Transplantation 2005; 80 (2(S)): 254–264. 16 Duncan N, Craddock C. Optimizing the use of cyclosporin in 34 Tremblay F, Fernandes M, Habbab F, de Edwardes M, allogeneic stem cell transplantation. Bone Marrow Transplant Loertscher R, Meterissian S. Malignancy after renal trans- 2006; 38: 169–174. plantation: incidence and role of type of immunosuppression. 17 Reynolds NJ, Al-Daraji WI. Calcineurin inhibitors and Ann Surg Oncol 2002; 9: 785–788. sirolimus: mechanisms of action and applications in dermato- 35 Herman M, Weinstein T, Korzets A, Chagnac A, Ori Y, Zevin logy. Clin Exp Dermatol 2002; 27: 555–561. D et al. Effect of cyclosporin A on DNA repair and cancer 18 Grawunder U, Zimmer D, Fugmann S, Schwarz K, incidence in kidney transplant recipients. J Lab Clin Med 2001; Lieber MR. DNA ligase IV is essential for V(D)J recombina- 137: 14–20. tion and DNA double-strand break repair in human precursor 36 Herman M, Ori Y, Chagnac A, Weinstein T, Korzets A, lymphocytes. Mol Cell 1998; 2: 477–484. Zevin D et al. DNA repair in mononuclear cells: role of serine/ 19 Riballo E, Kuhne M, Rief N, Doherty AJ, Smith GC, threonine phosphatases. J Lab Clin Med 2002; 140: 255–262. Recio MJ et al. A pathway of double strand break rejoining 37 Yarosh DB, Pena AV, Nay SL, Canning MT, Brown DA. dependent upon ATM, Artemis and proteins locating to Calcineurin inhibitors decrease DNA repair and apoptosis g-H2AX foci. Mol Cell 2004; 16: 715–724. in human keratinocytes following ultraviolet B irradiation. 20 Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM. J Investig Dermatol 2005; 125: 1020–1025. DNA double-stranded breaks induce histone H2AX phos- 38 O’Donovan P, Perrett CM, Zhang X, Montaner B, Xu Y-Z, phorylation on serine 139. J Biol Chem 1998; 273: 5858–5868. Harwood CA et al. Azathioprine and UVA light generate 21 Rogakou EP, Boon C, Redon C, Bonner WM. Megabase mutagenic oxidative DNA damage. Science 2005; 309: chromatin domains involved in DNA double-strand breaks 1871–1874. in vivo. J Cell Biol 1999; 146: 905–916. 39 Karran P. Thiopurines, DNA damage, DNA repair and 22 Fernandez-Capetillo O, Lee A, Nussenzweig M, Nussenzweig A. therapy-related cancer. Br Med Bull 2006; 79-80: 153–170. H2AX: the histone guardian of the genome. DNA Repair 2004; 40 Franco S, Gostissa M, Zha S, Lombard DB, Murphy MM, 3: 959–967. Zarrin AA et al. H2AX prevents DNA breaks from progres- 23 Fernandez-Capetillo O, Chen HT, Celeste A, Ward I, sing to chromosome breaks and translocations. Mol Cell 2006; Romanienko PJ, Morales JC et al. DNA damage-induced 21: 201–214. G2-M checkpoint activation by histone H2AX and 53BP1. Nat 41 Franco S, Alt FW, Manis JP. Pathways that suppress Cell Biol 2002; 4: 993–997. programmed DNA breaks from progressing to chromosomal 24 Lindahl T. Instability and decay of the primary structure of breaks and translocations. DNA Repair 2006; 5: 1030–1041. DNA. Nature 1993; 362: 709–714. 42 Couedel C, Mills KD, Barchi M, Shen L, Olshen A, 25 Caldecott KW. XRCC1 and DNA strand break repair. DNA Johnson RD et al. Collaboration of homologous recombina- Repair 2003; 2: 955–969. tion and nonhomologous end-joining factors for the survival 26 Caldecott KW. Mammalian single-strand break repair: mechan- and integrity of mice and cells. Genes Dev 2004; 18: 1293–1304. isms and links with chromatin. DNA Repair 2007; 6: 443–453. 43 KolbHJ, Socie G, Duell T, Van Lint MT, Tichelli A, Apperley JF 27 Caldecott KW. Mammalian DNA single-strand break repair: et al. Malignant neoplasms in long-term survivors of bone an X-ra(y)ted affair. Bioessays 2001; 23: 447–455. marrow transplantation. Ann Intern Med 1999; 131: 738–744.

Bone Marrow Transplantation