Association of ASXL1 and RUNX1 Variants with Splenomegaly in Myelodysplastic Syndromes Based on Next-generation Sequencing and Computed Tomography Data: A Retrospective Study

Abstract

Although splenomegaly is typically uncommon in myelodysplastic syndromes (MDS), it is associated with reduced engraftment rates and poor survival outcomes. Despite its clinical significance, the incidence and genetic associations of splenomegaly in MDS remain understudied. To address this, we conducted a retrospective study of 27 patients with MDS at the Veterans Health Service Medical Center in South Korea. Based on computed tomography scan evaluation, splenomegaly was identified in 26% of patients with MDS, and significant associations with variants in ASXL1 (P=0.0089 for null and missense/in-frame variants) and RUNX1 (P=0.042 for null variants) were observed, suggesting that these variants are linked to an increased risk of splenomegaly. Notably, one patient with ASXL1 and TET2 variants developed severe splenomegaly (spleen size, 29 cm) following granulocyte colony-stimulating factor (G-CSF) treatment, requiring splenectomy. This case suggests a potential interaction between specific genetic variants and G-CSF sensitivity, potentially exacerbating splenomegaly. Our findings suggest that the incidence of splenomegaly in patients with MDS, including mild cases, is likely underestimated and that ASXL1 and RUNX1 variants increase the risk of splenomegaly. Furthermore, careful monitoring for the development of severe splenomegaly during G-CSF treatment may be warranted in genetically susceptible individuals with MDS.

MDS is a clonal hematologic malignancy with an incidence of up to two per 100,000 individuals in South Korea [1]. Although uncommon in MDS [2], splenomegaly is associated with low engraftment rates and poor survival [3]. The incidence of splenomegaly and its association with genetic variants in patients with MDS are understudied. We investigated genetic risk factors for splenomegaly in 27 patients with MDS. Furthermore, we discuss the case of a patient with MDS harboring ASXL1 and TET2 variants who underwent surgery because of substantial spleen enlargement (29 cm in size) following granulocyte colony-stimulating factor (G-CSF) administration. This case raised the question of whether specific genetic variants induce susceptibility to G-CSF in patients.

To identify the genetic variants associated with splenomegaly in MDS, we retrospectively reviewed data from all patients with MDS (N=37) who visited the Veterans Health Service (VHS) Medical Center, Seoul, Korea, during 2020–2024. Inclusion criteria were: a final diagnosis of MDS confirmed using bone marrow (BM) analysis, next-generation sequencing (NGS)-based molecular profile of BM or blood at diagnosis based on the WHO classification of MDS (4th edition) available, and spleen size assessed using computed tomography (CT) at MDS diagnosis or later. Patients with other hematologic malignancies, chronic liver diseases (including cirrhosis and hepatitis), infections, venous thrombosis, autoimmune disorders, or other conditions causing splenomegaly were excluded. Cases of clonal hematopoiesis with no definite MDS evidence were carefully reviewed and excluded from this study [4]. Of the 37 patients with MDS, 27 were included. Clinical features, including age, sex, complete blood cell count, spleen size, international prognostic scoring system (IPSS) score, revised IPSS (IPSS-R) score, cytopenia, dysplasia, and progression to AML (Supplemental Data Table S1), were collected. MDS was diagnosed based on BM analysis using the Oncomine Myeloid Research Assay (OMRA) on an Ion S5XL NGS system (both from Thermo Fisher Scientific, Waltham, MA, USA) and chromosome analysis. The OMRA covers complete exon regions of 17 genes, exonic mutation hotspots of 23 genes, and 29 fusion genes (Supplemental Data Table S2). We categorized variants as null (nonsense, frameshift, or splice site) or missense/in-frame variants according to their gene effect. Splenomegaly was defined as a spleen size >9.76 cm (i.e., the upper limit of normality) on CT [5]. Data were analyzed using R v.4.4.1 (R Development Core Team, 2024). Fisher’s exact test was used to assess associations between splenomegaly and clinical features or genetic variants (Table 1 and Supplemental Data Table S3). Results with P<0.05 were considered statistically significant. The study complied with the Declaration of Helsinki and was approved by the Institutional Review Board of the VHS Medical Center (IRB No. 2020-04-003). In this retrospective study, we used fully anonymized data, and informed consent was therefore waived.

Table 1. Clinical and molecular features of 27 patients with MDS.

Feature		Splenomegaly (+) (N=7)	Splenomegaly (–) (N=20)	P
		Cases, N (frequency)
Age (yrs)	60–69	0 (0%)	1 (5%)	0.31
	70–79	6 (86%)	10 (50%)
	80–89	1 (14%)	9 (45%)
Sex	Male	7 (100%)	18 (90%)	1
	Female	0 (0%)	2 (10%)
IPSS	Low (0)	0 (0%)	6 (30%)	0.26
	INT-1 (0.5–1.0)	6 (86%)	11 (55%)
	INT-2 (1.5–2.0)	0 (0%)	2 (10%)
	High (≥2.5)	1 (14%)	1 (5%)
IPSS-R	Very low (≤1.5)	0 (0%)	2 (10%)	0.19
	Low (≤3.0)	3 (43%)	10 (50%)
	INT (≤4.5)	0 (0%)	5 (25%)
	High (≤6.0)	1 (14%)	1 (5%)
	Very high (>6.0)	3 (43%)	2 (10%)
Number of cytopenia	1	2 (29%)	4 (20%)	0.47
	2	3 (43%)	3 (15%)
	3	2 (29%)	13 (65%)
Number of dysplasia	0 or 1	1 (14%)	4 (20%)	1
	2 or 3	6 (86%)	16 (80%)
ASXL1	Null variant*	3 (43%)	1 (5%)	0.042
ASXL1	Null variant* + missense/in-frame variant	4 (57%)	1 (5%)	0.0089
RUNX1	Null variant*	3 (43%)	1 (5%)	0.042
RUNX1	Null variant* + missense/in-frame variant	3 (43%)	4 (20%)	0.33
TET2	Null variant*	5 (71%)	6 (30%)	0.084
TET2	Null variant* + missense/in-frame variant	6 (86%)	8 (40%)	0.077
SRSF2	Null variant*	0 (0%)	0 (0%)	1
SRSF2	Null variant* + missense/in-frame variant	2 (29%)	2 (10%)	0.27
DNMT3A	Null variant*	1 (14%)	1 (5%)	0.46
DNMT3A	Null variant* + missense/in-frame variant	1 (14%)	3 (15%)	1
TP53	Null variant*	0 (0%)	0 (0%)	1
TP53	Null variant* + missense/in-frame variant	0 (0%)	4 (20%)	0.55
U2AF1	Null variant*	0 (0%)	0 (0%)	1
U2AF1	Null variant* + missense/in-frame variant	0 (0%)	3 (15%)	0.55
SF3B1	Null variant*	0 (0%)	0 (0%)	1
SF3B1	Null variant* + missense/in-frame variant	2 (29%)	5 (25%)	1
EZH2	Null variant*	0 (0%)	3 (15%)	0.55
EZH2	Null variant* + missense/in-frame variant	1 (14%)	3 (15%)	1
STAG2	Null variant*	0 (0%)	2 (10%)	1
STAG2	Null variant* + missense/in-frame variant	0 (0%)	3 (15%)	0.55
SH2B3	Null variant*	0 (0%)	1 (5%)	1
SH2B3	Null variant* + missense/in-frame variant	1 (14%)	3 (15%)	1

*Null variant includes nonsense, frameshift, and splicing variants.

Abbreviation: IPSS, International Prognostic Scoring System.

Among the 27 MDS patients recruited, seven (26%) had splenomegaly (SPM group), and 20 (74%) did not (non-SPM group). The mean age was 77.2 and 79.1 yrs in the SPM and non-SPM groups, respectively. This study included 25 men and two women. The median spleen size in the SPM and non-SPM groups was 12.9 cm (range, 11.1–29.0 cm) and <9.76 cm, respectively. G-CSF use before CT evaluation of splenomegaly was observed in one of seven patients (14%) in the SPM group and five of 20 (25%) in the non-SPM group. Survival analysis (log-rank test) results did not achieve significance, likely because of the limited sample size.

Among the 40 genes analyzed, somatic variants were frequently observed in TET2 (14/27, 52%), RUNX1 (7/27, 26%), SF3B1 (7/27, 26%), ASXL1 (5/27, 19%), and others (Table 1 and Supplemental Data Fig. S1). Somatic variants were more prevalent in the SPM group than in the non-SPM group, with notably higher frequencies in ASXL1 (57%, 4/7 vs. 5%, 1/20), RUNX1 (43%, 3/7 vs. 20%, 4/20), and TET2 (86%, 6/7 vs. 40%, 8/20). Variants in ASXL1 (P=0.042 for null variants; P=0.0089 for null and missense/in-frame variants) and RUNX1 (P=0.042 for null variants; P=0.33 for null and missense/in-frame variants) were significantly associated with splenomegaly. TET2 variants showed a weak association with splenomegaly (P=0.084 for null variants; P=0.077 for null and missense/in-frame variants) (Table 1). When applying the diagnostic criteria of the WHO 5th edition (modified MDS, MDS/MPN, and AML diagnostic criteria) instead of the 4th edition classification [6], one of the 27 patients fulfilled the criteria for chronic myelomonocytic leukemia (CMML) and did not present with splenomegaly (Supplemental Data Table S1). Even after excluding this patient, associations with null and/or missense/in-frame variants in ASXL1 and null variants in RUNX1 remained significant (P<0.05). Compared with the group without CT evaluation (N=10), in which one of 10 patients (10%) had variants in both ASXL1 and RUNX1, the group with CT evaluation (N=27) did not show significant differences in the variant frequency (Fisher’s exact test).

ASXL1 variants, inducing leukemogenesis through loss of polycomb repressive complex 2-mediated gene repression, were significantly associated with splenomegaly, consistent with findings in animal studies [7]. Mice with ASXL1 knockout or ASXL1 with a stop codon variant at Y588 developed various myeloid lineage hematologic cancers, including MDS (18%–72%), and exhibited splenic infiltration of myeloid cells [8, 9]. Splenic infiltration of myelocytes and monocytes contributes to splenomegaly in various hematological malignancies, including chronic myeloid leukemia and CMML [2, 10]. This finding suggests that ASXL1 variants in MDS patients may cause splenomegaly through splenic infiltration.

Null variants in RUNX1 were also significantly associated with splenomegaly, consistent with a murine study demonstrating that mice harboring RUNX1 nonsense variants exhibited expansion of the red pulp with splenomegaly, along with the MDS phenotype in most cases (80%) [11]. Additionally, these findings suggest that C-terminal truncation variants in RUNX1, which are often observed in patients with MDS, induce MDS progression with a slightly enlarged spleen by enhancing DNA binding via the Runt domain [11, 12]. This is consistent with our findings; variants affecting C-terminal deletions in RUNX1 were more prevalent in the SPM group (3/3, 100%) than in the non-SPM group (1/4, 25%) (Supplemental Data Table S1 and Supplemental Data Fig. S2B).

One patient with MDS harboring ASXL1 and TET2 variants developed splenomegaly after receiving G-CSF and eventually underwent splenectomy (Fig. 1D). The patient was initially referred for BM examination because of pancytopenia, and the findings showed normal cellularity but erythroid dysplasia, including features such as internuclear bridging and karyorrhexis (Fig. 1A–C). Chromosome analysis revealed a normal karyotype, and NGS identified c.1874G>A (p.R625Q) in ASXL1 (Supplemental Data Fig. S2A) and c.1517G>A (p.R506K) and c.3119T>G (p.L1040*) in TET2 (Supplemental Data Fig. S2C). The patient was diagnosed as having MDS-single lineage dysplasia with a low IPSS-R score based on the WHO MDS classification (4th edition). One year after the diagnosis, the patient’s cytopenia persisted. G-CSF was administered intravenously (300 µg/0.7 mL filgrastim), with eight doses over five months. During this period, absolute neutrophil count (ANC) increased from 400/µL to 700/µL (>1,800/µL) (Fig. 1E), and the spleen size was 23 cm on CT. Because of insufficient recovery from cytopenia, an additional 18 doses of G-CSF were administered over three months, resulting in a rapid increase in ANC to 13,390/µL. However, the patient presented with left upper quadrant discomfort and underwent splenectomy because of rapid spleen enlargement (29 cm, 3,450 g post-operatively). Post-surgery, the persistent thrombocytopenia resolved quickly, and the patient has since maintained regular outpatient follow-up, with no evidence of cytopenia. In this case, splenomegaly may have been associated with G-CSF administration and specific genetic factors.

Fig. 1. BM analysis and clinical course of a patient with MDS harboring ASXL1 and TET2 variants. (A, B) Erythroid dysplasia, including internuclear bridging (A) and karyorrhexis (B), was observed in BM aspirates with Wright–Giemsa stain (400×). (C) BM biopsy, stained with hematoxylin and eosin, showed normal cellularity for age (40×). (D) The markedly enlarged spleen was dark reddish-brown, indicative of congestion, measuring 29 cm and weighing 3,450 g. (E) Laboratory findings and clinical course, including Hb, platelet, WBC, and ANC levels over 700 days after MDS diagnosis. Black arrows above the upper panel indicate the spleen size and time of splenectomy; those below the lower panel indicate the days of granulocyte colony-stimulating factor administration.

Abbreviations: BM, bone marrow; WBC, white blood cell count; ANC, absolute neutrophil count.

Complications such as splenomegaly and extramedullary hematopoiesis are known consequences of G-CSF therapy, and massive red pulp congestion and splenic rupture in patients with MDS following G-CSF treatment have been reported anecdotally [13]. Considering that ASXL1 variants promoted myeloid infiltration in the spleen in mice with MDS, G-CSF may synergistically exacerbate myeloid cell infiltration in patients with MDS, potentially leading to splenomegaly [14]. Hypersensitivity to G-CSF has been observed in mice with a RUNX1 variant, although the genetic variant differed from that in our patient [15]. These findings suggest that G-CSF and specific gene variants can synergistically contribute to splenomegaly via various mechanisms.

We observed splenomegaly in over 25% of patients with MDS, along with significant associations between splenomegaly and the ASXL1 and RUNX1 variants. The prevalence of splenomegaly was substantially higher than previously reported [2]. However, selection bias was unlikely, as most patients with MDS (72.9%) were enrolled, and except for TET2 variants, clinical features did not differ significantly between recruited and non-recruited patients. This difference in splenomegaly prevalence suggests that mild splenomegaly may often be overlooked in clinical practice owing to the lower sensitivity of physical examination compared with that of imaging [16]. Further research is required to evaluate the prevalence of splenomegaly in MDS and confirm the association between ASXL1 and RUNX1 variants and splenomegaly, suggesting that splenic evaluation guided by genetic testing results should be considered. As fatal events, including splenic rupture, have been reported following G-CSF use in some patients [13], careful reconsideration of G-CSF use in patients with MDS, particularly in those with genetic susceptibilities, is necessary. Large-scale studies are required to establish the clinical implications of these findings before practical implementation.

SUPPLEMENTARY MATERIALS

Supplementary materials can be found via https://doi.org/10.3343/alm.2025.0033.