Granulocyte–Colony-Stimulating factor (G-CSF) is currently the standard mobilising agent for peripheral blood stem cell (PBSC) donation. Concerns that it may trigger chromosome aberrations similar to those observed in leukaemia patients were refuted but long-term effects of G-CSF mobilisation on genome integrity remains unclear. In the setting of a multi-centre clinical trial we screened blood samples from 50 PBSC donors at cellular and gene level for aberrations common in haematological malignancies using fluorescence in situ hybridisation (FISH) and next generation sequencing (NGS) assays. Analysis of samples collected before, on the day of donation, 90 and 180 days after G-CSF admission confirmed the absence of short-term effects in PBSC donors on both quiescent and dividing cells. This data did not differ from the results of 50 individuals tested 3–5 years after bone marrow donation and 50 healthy persons. NGS using a panel targeting 54 genes recurrently affected in myeloid disorders (TruSight Myeloid panel, Illumina) showed that the gene profiles of samples from 48 PBSC donors remained stable throughout the study period. These data strongly indicate absence of detrimental effects on the genome integrity caused by PBSC donation.
Granulocyte–Colony-Stimulating Factor (G-CSF) is a hematopoietic cytokine produced by monocytes, fibroblasts and endothelial cells whose principal role in normal steady-state haematopoiesis is to regulate production, differentiation and functional activation of neutrophils along with a number of other biologically important activities . Recombinant G-CSF has been used for over two decades in the management of chronic idiopathic neutropenia, chemotherapy-induced myelosuppression and to accelerate neutrophil recovery after allogeneic or autologous stem cell transplantation [2,3,4]. G-CSF is also used as the standard mobilising agent for peripheral blood stem cell (PBSC) donation  and is licensed for this purpose.
While G-CSF use is associated with mild transient side-effects, concerns that it may trigger haematological malignancy were raised by a study of PBSC donors found to display epigenetic and genetic changes similar to those observed in leukaemia patients . Meanwhile, a substantial amount of prospective and retrospective data has accumulated to allow assessment of the incidence of myeloid malignancies after G-CSF administration [5, 7]. Three AML cases (0.006%) and 20 other haematological malignancies (0.04%) were identified in a total of 44,716 donors included in 12 studies followed for a total of nearly 170,000 donor-years. These data clearly demonstrate that the frequency of AML in G-CSF treated individuals is not notably different from that reported in the general population .
Although several studies have failed to reproduce the original data, possible long-term effects of G-CSF mobilisation on chromosome integrity remain unclear [9, 10]. In the setting of a multi-centre clinical trial we screened by fluorescence in situ hybridisation (FISH) selected haematological significant genome loci and confirmed the absence of long-term effects in balanced prospective and retrospective arms of the study. However, human malignancies are characterised by a spectrum of genome aberrations—from large scale chromosome changes to sub-microscopic alterations and single nucleotide variations . Recent advances in high-throughput techniques, referred to as next generation sequencing (NGS) enable global genome assessment which allow identification of genomic profiles of many tumour types thereby reshaping our understanding of malignant disease [12,13,14]. Screening for selected gene variance signatures in peripheral blood leucocytes by NGS offers a highly sensitive way of detecting genome damage that may lead to malignant disorder.
The aim of our trial was to address the limitations of previous attempts to evaluate the possible long-term effect of G-CSF mobilisation on chromosome integrity in PBSC donors. In addition, we screened for selected gene variance signatures in peripheral blood leucocytes of G-CSF mobilised donors for changes observed in patients with haematological malignancies. The data obtained by both FISH and NGS disprove the claims that G-CSF mobilisation may cause genetic and epigenetic changes.
Materials and methods
This study was conducted as a Clinical Trial under the governance of the UCL Trial Office in compliance with MHRA and IRB requirements (REC reference numbers 08/H0720/2 and 16/WS/0036; EudraCT number: 2007-006301-24 and IRAS project ID:198259). All 200 participants, recruited in three different hospitals, were consented and none had a family history of haematological malignancies. We screened samples by FISH to assess aneuploidy and structural chromosome aberrations as follows:
A prospective arm of 50 unrelated PBSC donors, 41 males and 9 females (average age 36 years) were given a standard dose of G-CSF (Lenograstim) at 10 ug/kg (rounded to the nearest vial size) for three consecutive days. Peripheral blood samples were obtained at four time points i.e., before G-CSF administration; on the day of donation; 90 days and 180 days post G-CSF administration. A full set of four samples was available for 45 of the 50 donors.
A retrospective arm including 50 of each: (i) PBSC donors, 3–5 year post G-SCF mobilisation (41 males and 9 females, average age 41 years), (ii) Bone marrow (BM) donors (29 males and 21 females, average age 39.7 years), 3–5 years post donation and (iii) healthy control subjects, not pregnant and without family history of cancer or prior exposure to G-CSF, demographically matching the tested group (36 males and 14 females, average age 41 years).
Dual colour/dual probe FISH assay
We replicated the conditions of the Nagler et al. (2004) study  by using cells from PHA stimulated peripheral blood cultures while FISH was carried out as before  (see Supplementary Notes). Dual colour/dual probe FISH sets (D-FISH) were made using BAC clones verified on normal human metaphase cells. The D-FISH probe sets for chromosomes 7, 8 and 17 consist of target sequences labelled in red and control in green with genome address, size and chromosome location of all BAC probes listed in Supplementary Table 1a. All samples were coded on collection and blindly screened. For FISH scoring and statistical methods see Supplementary notes.
After the FISH studies were completed gene profiling was carried out on 48 stem cell donors following a protocol used to screen over 3000 samples to date suspected or known to have haematological malignancies. Genomic DNA samples were obtained from 48 donors. The TruSight Myeloid target gene panel was read on a MiSeq platform (Illumina, San Diego, Ca 92122, USA) to screen mutational hotspots in 54 cancer-related genes relevant in myeloid malignancy (Supplementary Notes). Exonic and non-synonymous gene variances were reported at allele frequencies (VAF) ≥ 5% and at minimum read depth of 300 as per manufacturers criteria. We used the Catalogue of Somatic Mutations In Cancer (COSMIC), dbSNP and 1000 genome (≤2%) to classify gene variants as either drivers, variants of unknown significance and germline single nucleotide polymorphisms (SNPs). Genome addresses are given according to hg19 (GRCH37).
Gene variant profiles
Altogether 128 samples from 48 stem cell donors were screened for target gene variants using the TruSight myeloid panel (Table 1). In 9 out of the 48 donor’s DNA samples collected before and post donation (90 or 180 days) were investigated, while the rest had a third sample from the apheresis also screened for gene variants.
(i) Common variants
In 22 out of the 48 donors (48%) no gene variants were detected at levels above 5% allele frequency while a total of 42 variants were identified in the remaining 26 donors (Table 1). In 15 out of the 54 gene regions targeted by the Myeloid panel total of 32 different gene variants were identified. These include eight different variants of TET2, four of ASXL1 and CSF3R, two of BCOR, RUNX1, ETV6 and CUX1 while one variant was detected in the remaining nine genes in this sample cohort (Fig. 1a and Table 1). The vast majority (19/32) of these gene variants are reported as SNP in apparently healthy individuals, while further eight were also reported in cancer patients but considered to be tolerated or without known clinical significance. The remaining gene variants namely ASXL1p.Glu1102Asp and ASXL1p.E1102D, TET2p.Tyr867His, BCOR1p.S1223L and RUNX1p.L56S have a dual history i.e., detected in myeloid neoplasms (MDS, CMML, AML and MF) as well as reported as benign variant in healthy persons (http://cancer.sanger.ac.uk and http://www.ncbi.nlm.nih.gov/SNP). Gene variants were identified at either near haploid 47–55% or 98–99% variant allele frequency (VAF) (Table 1).
(ii) Dynamics of gene variant profile
Three samples, i.e., before donation, at the time of apheresis and post G-CSF mobilisation (90 or 180 days) were investigated in 39/48 donors, a further 9 donors had two samples screened (before and post G-CSF mobilisation at 90 or 180 days). In half of the donors harbouring gene variants, only one variant was detected, 8 had two, a further 4 had three and only one (donor 5, female age 20) had 5 variants (Table 1). The gene profiles remain stable during the period of the study in all but four donors (Fig. 1b, Table 1). The latter shared a common feature—loss of one gene variant in the sample collected 180 days post GCSF mobilisation, namely donors 8 and 12 showing loss of CUX1 and donors 21 and 25—loss of ASXL1 (Table 1 and Fig. 1c). In contrast, gains of any novel gene variants were not observed in this sample cohort.
The number of aberrant cells recorded by interphase (iFISH) and metaphase (mFISH) are summarised in Supplementary Tables 2-4, while the assessment of false signal range for the dual colour/dual set probes is presented in Supplementary Table 1b. Of the 352 samples received, only four (1.2%) failed to yield metaphases for analysis: two PBSC donors, one BM donor and one healthy control. In these samples FISH screening was carried out solely on interphase cells. No abnormalities were seen above the threshold for a positive result in any of the groups (Supplementary Fig. 1). The number of aberrant cells found in each individual together with the standard error of the mean (SEM), overall totals and representative images are shown in Supplementary Tables 2-4.
In an attempt to finally resolve the residual uncertainty surrounding the effects of the G-CSF administration we used NGS to search for cryptic genome aberrations that may remain undetected by FISH screening.
Our study differs from the previous reports in several ways. Firstly, we extended the analysis regime to NGS and secondly, we assessed both the short-term (prospective arm) and long-term (retrospective arm) effect of G-CSF mobilisation. In contrast to any previous studies we also searched for aneuploidies in BM donors 3–5 years post donation, demographically matching the PBSC group (positive controls).
It was possible to screen DNA obtained from 48 out of the 50 G-CSF mobilised stem cell donors for gene mutations using target NGS panel and compare these results with the comprehensive FISH data. The NGS analysis represents an almost two orders of magnitude increase in coverage for seeking genomic damage compared with the earlier chromosome and FISH studies.
Genome-sequencing studies indicate that all humans carry many genetic variants predicted to cause loss of function of protein-coding genes, one variant every eight bases of the exome the majority of which (>75%) have no currently established human disease phenotype . Accumulation of such gene variants is a recognised age-related phenomenon  but also a hallmark of clonality, which defines any neoplasia. Referred to as clonal haematopoiesis of undetermined origin (CHIP) in middle age and older people, this phenomenon includes a handful of genes, namely DNMT3A, TET2, and ASXL1 that encode epigenetic modifiers . However, the progression into overt neoplasia involves the accumulation of multiple mutations that arise from unavoidable errors associated with DNA replication and/or environmental (including drug therapy) and inherited factors [19,20,21].
As acquisition of mutations is not likely to be completely random, we set out to search for novel or modified CHIP variants in stem cell donors that could be linked to the G-CSF mobilisation. The gene profiles of samples collected before, at apheresis and after donation were found to remain essentially constant indicating lack of clonal evolution. Although not considered to be a marker for malignant disease the same ASXL1 (3’UTR variant), rs184265056 was missing in the p180 days post donation sample of one donor, whilst present at haploid levels before donation and in the apheresis samples (Table 1). We detected several variants, notably ASXL1p.Glu1102Asp (c.3306G > T) as well as TET2p.Tyr867His, RUNX1p.L56S and BCORp.S1189L that are reported to occur in patients with myeloid disorders as well as in healthy individuals. For donors in our cohort, they were present prior to GCSF administration and their allele frequency remained constant for the 180 days.
The dynamics of these gene profiles differ from one that we have observed in over 3000 samples from patients with myeloid disorders using the same TruSight Myeloid panel where increase of frequency and/or acquisition of novel gene mutations marks disease progression revealed by clinical and laboratory findings . While we expect to find clonal haematopoiesis in some donors a comparison of the gene variance profile of samples before and after stem cell donation is evidence that no damage occurs.
We screened both quiescent and dividing cells for aberrations at frequency levels well below the threshold for a positive FISH dual colour/dual probe assay. We chose to screen loci most commonly affected in de novo and secondary AML/MDS. In addition to the ubiquitous TP53 gene, prevalent in advanced stages of tumour progression, we included the MYC gene at chromosome 8q24 along with two sites of frequent deletions in the long arm of chromosome 7. None of the aneuploidies observed in our FISH tests can be interpreted to suggest a pathological condition since they were detected in single cells and likely to represent apoptotic cells or experimental artefacts. They also exist at similar levels in both sets of controls and mobilised donors. Furthermore, metaphase cell analysis of the same samples provided no evidence for clonal chromosome abnormalities, the hallmark of malignant proliferation. In the absence of a clonal expansion, cells with t(7;14)(q34;q11) seen as a single event in six subjects represent an established phenomenon related to the PHA stimulation . Similarly, the presence of limited number of cells with polyploidy are associated with in vitro cell culturing.
The FISH findings from both prospective and retrospective arms of our study are consistent with other published data [9, 10], and clearly demonstrate the absence of a short and long-term detrimental effect on chromosome integrity caused by G-CSF administration in PBSC donors. We are aware that the sample size of our NGS investigations is too small to allow definitive conclusions regarding possible gene damage incurred by the GSCF. However, we searched and failed to record any changes in the gene profile before and post donation in all donors that are suggestive for clonal expansion as seen in haematological malignancies. Undoubtedly further studies are required to fully assess the possible role of G-CSF mobilisation on gene integrity and especially in stem cell donors who carry CHIP type gene mutations. As massive gene variance screening is already part of the diagnostic routine in onco–haematology it would be prudent to be included in the selection procedure of stem cell donors. The observation provided by our study that gene profiles obtained with target TruSight Myeloid panel (Illumina) remain stable 180 days post donation strongly indicates that G-CSF mobilisation of stem cell donors is unlikely to confer genetic mutations.
Anderlini P, Przepiorka D, Champlin R, Körbling M. Biologic and clinical effects of granulocyte colony-stimulating factor in normal individuals. Blood 1996;88:2819–25.
Pamphilon D, Nacheva E, Navarrete C, Madrigal A, Goldman J. The use of granulocyte-colony-stimulating factor in volunteer unrelated hemopoietic stem cell donors. Transfusion 2008;48:1495–501.
Battiwalla M, McCarthy P. Filgrastim support in allogeneic HSCT for myeloid malignancies: a review of the role of G-CSF and the implications for current practice. Bone Marrow Transpl. 2009;43:351–6.
Beekman R, Touw IP. G-CSF and its receptor in myeloid malignancy. Blood 2010;115:5131–6.
Martino M, Fedele R, Massara E, Recchia A, Irrera G, Morabito F. Long-term safety of granulocyte colony-stimulating factor in normal donors: is it all clear? Expert Opin Biol Ther. 2012;12:609–21.
Nagler A, Korenstein-Ilan A, Amiel A, Avivi L. Granulocyte colony-stimulating factor generates epigenetic and genetic alterations in lymphocytes of normal volunteer donors of stem cells. Exp Hematol. 2004;32:122–30.
Avalos BR, Lazaryan A, Copelan EA. Can G-CSF cause leukaemia in hematopoietic stem cell donors? Biol Blood Marrow Transpl. 2011;17:1739–46.
Shaw B, Confer D, Hwang W, Pulsipher M. A review of the genetic and long-term effects of G-CSF injections in healthy donors: a reassuring lack of evidence for the development of haematological malignancies. Bone Marrow Transpl. 2015;50:334–40.
Marmier-Savet C, Larosa F, Legrand F, Witz B, Michallet M, Ranta D, et al. G-CSF-induced aneuploidy does not affect CD34+ cells and does not require cell division. Blood 2010;115:910–1.
Hirsch B, Oseth L, Cain M, Trader E, Pulkrabek S, Lindgren B, et al. Effects of granulocyte-colony stimulating factor on chromosome aneuploidy and replication asynchrony in healthy peripheral blood stem cell donors. Blood 2011;118:2602–8.
Mitelman F, Heim S. Cancer Cytogenetics: chromosomal and molecular genetic aberrations of tumour cells. 4th Edition. Wiley-Blackwell. 2015. p.1–10.
Forbes SA, Bindal N, Bamford S, Cole C, Kok C, Beare D, et al. COSMIC: mining complete cancer genomes in the Catalogue of Somatic Mutations in Cancer. Nucleic Acids Res. 2011;39(supl. 1):D945–50.
Network C, Weinstein JN, Collisson EA, Mills GB, Shaw KR, Ozenberger BA, et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet. 2013;45:1113–20.
Nangalia J, Mitchell E, Green AR. Clonal approaches to understanding the impact of mutations on hematologic disease development. Blood 2019;133:1436–45.
Virgili A, Nacheva EP. Genomic amplification of BCR/ABL1 and a region downstream of ABL1 in chronic myeloid leukaemia: a FISH mapping study of CML patients and cell lines. Mol Cytogenet. 2010;3:15.
Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fennell T, et al. Analysis of protein-coding genetic variation in 60,706 humans. Nature 2016;536:285–91.
Stratton MR. Journeys into the genome of cancer cells. EMBO Mol Med. 2013;5:169–72.
Steensma DP. Clinical Implications of Clonal Haematopoiesis. Mayo Clin Proc. 2018;93:1122–30.
Bodian DL, McCutcheon JN, Kothiyal P, Huddleston KC, Iyer RK, Vockley JG, et al. Germline variation in cancer-susceptibility genes in a healthy, ancestrally diverse cohort: implications for individual genome sequencing. PloS one. 2014;9:e94554.
Sperling AS, Gibson CJ, Ebert BL. The genetics of myelodysplastic syndrome: from clonal haematopoiesis to secondary leukaemia. Nat Rev Cancer. 2017;17:5–19.
Tomasetti C, Li L, Vogelstein B. Stem cell divisions, somatic mutations, cancer aetiology, and cancer prevention. Science. 2017;355:1330–4.
Boneva T. Prognostic significance of gene mutations in MDS depends on the loci of gene variances. EHA Library 2017;180952. (Abstract E1176).
Dewald G, Noonan K, Spurbeck J, Johnson D. T-lymphocytes with 7;14 translocations: frequency of occurrence, breakpoints, and clinical and biological significance. Am J Hum Genet. 1986;38:520–32.
We are grateful to all the donors who agreed to take part in this study and the nursing staff who assisted us. We also thank all the office staff at the British Bone Marrow Registry (BBMR) and at Anthony Nolan, for their support in tracing the samples. The laboratory investigations were performed by A. Virgili, K. Gancheva, L. Rai, G. Uka and S. Phyo. Chugai, Amgen and Illumina, the Royal Free Charity, the BBMDA Charity, Anthony Nolan and the BBMR through NHSBT provided the financial support for this study.
Conflict of interest
The authors declare that they have no conflict of interest.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Nacheva, E., Ahyee, T., Addada, J. et al. Absence of damaging effects of stem cell donation in unrelated donors assessed by FISH and gene variance screening. Bone Marrow Transplant 55, 1290–1296 (2020). https://doi.org/10.1038/s41409-020-0945-y