Impact of <i>CHEK2</i> germline variants on haematological malignancy risk and outcomes of allogeneic <scp>HSCT</scp>

The significance of germline genetics in haematological malignancies (HM) is increasingly acknowledged. This has led to the development of clinical guidelines for recognizing the risk of hereditary HMs and in selected patients tailoring the conditioning and graft-versus-host disease (GVHD) prophylaxis undergoing haematopoietic stem cell transplantation (HSCT).1-7 Today, more than 100 genes associated with increased risk for HM are acknowledged and included in the germline screening panels. Predisposition to cancer is most often caused by aberrations in genes involved in key cellular functions, such as DNA damage repair (DDR). Variants in CHEK2, a DDR gene, predispose to solid cancer, particularly breast cancer.8-10 Recent data suggest that rare germline CHEK2 variants may also affect the risk of myeloid HMs.11, 12 Furthermore, we have observed an increased risk for acute lymphoblastic leukaemia.13 The most studied CHEK2 variants globally, (NM_007194.4) c.1229del, p. Cys410SerfsTer4 and c.470T>C, p. Ile157Thr, show a markedly higher prevalence in the Finnish population (0.86% and 2.58%, respectively) compared to non-Finnish populations (0.19% and 0.18%, respectively).14 The ascending role of CHEK2 variants in HM predisposition and the lack of reports on their impact on the outcomes of HSCTs prompted us to perform this study. The patients and analysis methods are described in Figure 1A; Supporting Information and Table S1. We analysed data from adult and paediatric patients who underwent HSCT (n = 877), along with a subset of paired HSCT donors (n = 669). The study cohort comprised three adult and one paediatric cohort (Figure 1A). We used whole exome sequencing (WES) data of patient samples and analysed germline variants with a panel of genes predisposing to haematological malignancies (haematology panel) and solid cancers (oncology panel) as previously described in Lahtinen et al4 (Table S2). To validate our findings, we used an independent HSCT cohort of patients with single nucleotide polymorphism (SNP) array data covering the CHEK2 c.1229del and c.470T>C (Figure 1A: Adult cohort 3). Furthermore, we used WES and SNP array data available for the respective HSCT donors (Figure 1A; Table S3). During data analysis, we focused on variants that can be reliably analysed using WES and SNP methods. However, we acknowledge that the detection of intronic variants is a limitation of our study design. We analysed overall survival (OS), relapse rates and non-relapse mortality (NRM) using Kaplan–Meier estimations and log-rank tests. Five patients of 391 with multiple variants in the WES data and three patients carrying two CHEK2 variants identified in the SNP cohort were excluded from the analyses. Differences in minor allele frequencies (MAF) between patients, donors and the gnomAD Finnish non-cancer controls (gnomAD v4.1.0 Finns, n = 32 026)14 were assessed using two-sided Fisher's exact tests. Given the genetic isolation of the Finnish population, gnomAD provides an adequately powered reference cohort for comparison. The 877 patient samples provided at least 80% power to detect a twofold increased risk of HMs associated with CHEK2 variants. Germline pathogenic (P) or likely pathogenic (LP) variants were identified in 3.6% (haematology panel, n = 14/386) and 6.2% (oncology panel, n = 24/386) of the patients in the WES dataset (Table S4). The most prevalent P/LP variants found in the genes included in the panels were in CHEK2. Thus, we also included CHEK2 c.470C>T in further analyses considering its association with cancer risk, despite having conflicting classifications of pathogenicity.13 Clinical parameters did not significantly differ between the CHEK2 variant, haematology panel and oncology panel variant carriers (Table S5). The proportion of family donors was slightly lower in patients with a gene variant included in the haematology panel compared to the CHEK2 variant, oncology panel variant and no variant groups (35.7% vs. 56.8%, 58.3% and 52.3%, respectively) likely reflecting the inherited nature of the disease (Table S5). The median follow-up time after HSCT was 46 months (range: 0–258 months). The OS, relapse rate and NRM among patients with the CHEK2 or oncology panel variants did not differ from those without an identified variant. In contrast, patients carrying P/LP variants included in the haematology panel had a significantly lower OS (p = 0.021) compared to patients without identified germline variants (Figure 1B). The difference in the OS resulted from a combination of excess NRM and increased relapse rate, but these findings were not significant (Figure S1). Table S6 presents demographic data, transplant-related characteristics and the outcome of patients carrying a P/LP variant in the haematology panel. We did not detect differences in rates of acute (severe or all stages) or chronic (extensive or all stages) GVHD between any of the groups (Figure S2). When analysing the SNP cohort, we found that 4.9% (n = 24/486) and 9.7% (n = 47/486) of patients carried c.1229del and c.470C>T, respectively. We then combined the WES and SNP datasets for further analyses. In total, we identified the CHEK2 c.1229del variant in 36/877 (4.1%; 2 homozygous and 34 heterozygous) and c.470T>C in 82/877 (9.4%; 5 homozygous and 77 heterozygous) patients (Table 1). The MAFs were significantly higher in patients than in the gnomAD Finnish population control group: 0.022 versus 0.009 (c.1229del, p < 0.001) and 0.050 vs. 0.025 (c.470T>C, p < 0.001). Remarkably, the odds ratio for our patients with HM was 2.5 (95% CI, 1.7 to 3.5) and 1.9 (95% CI, 1.5 to 2.4) for c.1229del and c.470T>C respectively (Table 1). Among acute leukaemia patients, the prevalence of the CHEK2 c.1229del variant was similar in paediatric and adult groups (OR 2.73, 95% CI 1.11–6.73; and OR 2.54, 95% CI: 1.57–4.10 respectively). In contrast, the MAF of the c.470C>T variant was lower in paediatric patients (OR 1.27, 95% CI 0.59–2.73) compared to adults (OR 2.15, 95% CI: 1.57–2.95), suggesting a potentially weaker association in children. Due to the limited paediatric sample size, we refrain from making strong conclusions. A further analysis was conducted on patients carrying either CHEK2 c.1229del (n = 33) or c.470T>C (n = 79) variants (three patients carrying both were excluded from the analyses). We detected no difference in OS, relapse rate, NRM or GVHD either within the CHEK2 variant groups or compared to patients without either variant (Figure 1C; Figures S2 and S3). Both patients homozygous for the CHEK2 c.1229del variant died of non-relapse-related causes within the first year after HSCT. Two of the five patients homozygous for the c.470C>T variant relapsed, leading to the death of one of them, while four survived until the end of the follow-up period with a median follow-up time of 5.2 years. Seventeen (2.5%; 1 homozygous and 16 heterozygous) and 47 (7.0%; 2 homozygous and 45 heterozygous) of the 669 donors carried CHEK2 variants c.1229del and c.470T>C respectively (Table S3). The prevalence of donors harbouring the CHEK2 variants, some of whom were related to patients, exceeded that observed in the Finnish control population but fell below the prevalence noted in our patient cohort (Table 1). We did not find a significant impact on the recipient OS, relapse rate or NRM when comparing donors with the CHEK2 variants to those with CHEK2 wild type (Figure S4). In addition to CHEK2 variants c.1229del and c.470T>C, we found three variants in the haematology panel in three donors and two variants in the oncology panel in six donors (Table S7). These variants had been identified in the respective HSCT recipients. All three patients who received transplants from homozygous CHEK2 carriers (one c.1229del and two c.470C>T) were alive at the end of follow-up, with a median follow-up time of 6.5 years with one experiencing relapse during that time. Patients' or donors' CHEK2 variants did not have a significant impact on the risk of death, relapse rate or non-relapse mortality in these analyses (Figure S4). The results of the multivariate survival analysis were consistent with the acknowledged risk factors associated with HSCT outcomes (Table S8). High-quality care for patients with HM involves comprehensive genetic analyses and the integration of findings into clinical practice. Here, we focused on patients harbouring germline CHEK2 variants who had undergone allogeneic HSCT due to HM. Our study consisted of 877 HSCT recipients including 118 with germline CHEK2 variants. By comparing with population-matched controls, we demonstrate that carriers of the CHEK2 c.1229del or c.470T>C variants have a twofold increased risk of developing HMs requiring HSCT. These findings further support the role of CHEK2 as one of the most prevalent inherited risk factors for HM.11-13 CHEK2 carriership did not significantly affect 5-year post-HSCT outcomes, including overall survival, the incidence of GVHD or NRM. This contrasts with findings for other DDR genes, such as bi-allelic mutations in Fanconi anaemia genes or DDX41, which have been associated with elevated risks of post-transplant toxicity and acute GVHD.5-7 However, homozygosity for c.1229del may confer an increased risk of adverse post-HSCT outcomes, which is shown by the early mortality observed within the first year following transplantation in both patients carrying the homozygous variant. Although short-term outcomes following HSCT were favourable for heterozygous CHEK2 carriers, these variants may still predispose individuals to an elevated risk of secondary malignancies with prolonged latency periods. Given that CHEK2 is primarily associated with malignancies typically presenting in adulthood,8-10 prior findings emphasize the importance of long-term post-HSCT surveillance to assess the potential impact of CHEK2 variants on late-onset complications and secondary cancer development. Importantly, although CHEK2 variant carriership was associated with an increased predisposition to high-risk HMs, a HSCT from a donor harbouring these variants, including homozygous carriers, did not negatively affect HSCT outcomes over the 5-year follow-up period. This suggests that stem cell donors harbouring CHEK2 variants may have a comparatively smaller impact than high- or intermediate-risk cancer predisposing gene variants such as DDX41, GATA2 and RUNX1.6, 15 In conclusion, our findings exemplify the need for gene-specific and possibly even variant-specific studies aimed at the optimal integration of germline genetics into haematology clinical practice. These authors contributed equally: Atte K. Lahtinen and Maarja Karu. These authors contributed equally: Kirsi Jahnukainen, Outi Kilpivaara & Ulla Wartiovaara-Kautto. Atte K. Lahtinen analysed data and performed statistical analysis; Maarja Karu analysed data; Jarmo Ritari, Kati Hyvärinen, Satu Koskela, Julia Nihtilä, Jukka Partanen, Kim Vettenranta, Minna Koskenvuo, Riitta Niittyvuopio, Urpu Salmenniemi and Maija Itälä-Remes contributed data; Kirsi Jahnukainen, Ulla Wartiovaara-Kautto and Outi Kilpivaara designed the study, supervised students, and interpreted the data; Atte K. Lahtinen, Maarja Karu, Kirsi Jahnukainen, Ulla Wartiovaara-Kautto and Outi Kilpivaara wrote the manuscript. All authors contributed to the critical review of the manuscript and have approved the manuscript. Open access publishing facilitated by Helsingin yliopisto, as part of the Wiley - FinELib agreement. Lotta Katainen and Marja Pekkanen are thanked for their extremely precise technical help. The Finnish Hematology Registry and Clinical Biobank provided some of the samples and data used in this project. This work was supported by the Research Council of Finland #349760, Cancer Foundation Finland, Sigrid Jusélius Foundation, the Finnish Special Governmental Subsidy for Health Sciences, Research, and Training, the Helsinki University Hospital Comprehensive Cancer Research Funding, the Finnish Funding Agency for Technology and Innovation (TEKES), the Väre Foundation for Pediatric Cancer Research, the iCAN Digital Precision Cancer Medicine, Cancer Foundation Finland, Ane and Signe Gyllenberg Foundation, foundation for Pediatric Research, the Swedish Childhood Cancer Foundation (KP2020-0012) and the Birgitta and Carl-Axel Rydbeck's Research Grant for Paediatric Research (2020-00335, 2021-00079 and 2023-00380). We would also like to acknowledge support by the Finnish Medical Foundation, the Finnish Association of Hematology, the Ida Montin Foundation, the Biomedicum Helsinki Foundation and Jalmari and Rauha Ahokas Foundation. The authors declare no conflicts of interest. Samples and data were collected after obtaining informed consent from study participants or authorization by the ethics committee for deceased patients. The study has been approved by the Ethics Review Boards of each collaborating hospital and the Finnish National Supervisory Authority for Welfare and Health (Valvira). The permit numbers are #206/13/03/03/2016 (amendment 1Q/2023), #303/13/03/01/2011, HUS/114/2018, HUS/284/2019 and V/3235/2019 for the WES set and V/74832/2017, HUS/2152/2020 and ETMK 78/2012 for the SNP array set. For original, de-identified data, please contact the corresponding authors. Data S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.