Analysis of the 3′UTR region of the NOTCH1 gene in chronic lymphocytic leukemia patients
Summary. Deregulation of NOTCH1-signalling pathway is common in chronic lymphocytic leukemia (CLL). The most of studies are focused on detection of the hotspot c.7541_7542delCT NOTCH1 mutations in exon 34, while studies of mutations in the 3′UTR region are rare. The aims of work were to evaluate the frequencies of mutations in the 3′UTR region of the NOTCH1 gene (9:136,495553-136,495994) in Ukrainian CLL patients, the distribution of rs3124591 genotypes located in that area, and association of NOTCH1 mutations with structure of B-cell receptor. Materials and Methods: Detection of mutations in the 3′UTR region of the NOTCH1 was performed by direct sequencing in 87 previously untreated CLL patients (from the total group of 237 CLL patients) with unmutated immunoglobulin heavy-chain variable (UM IGHV) genes and without mutations in hotspot regions of TP53, SF3B1, and exon 34 of NOTCH1 genes. Results: Mutations in the 3′UTR region of the NOTCH1 were revealed in three of 87 CLL patients (3.4%). Two cases with non-coding mutations were related to subset #1 of stereotyped B-cell receptors, and one case belonged to stereotyped subset #28a. Analysis with inclusion of 30 UM IGHV cases with previously detected c.7544_7545delCT mutations revealed that the frequency of UM IGHV genes of I phylogenetic clan (except IGHV1-69) was significantly increased, and the frequency of UM IGHV3 and IGHV4 genes, on the contrary, was reduced in NOTCH1-mutated cases comparing with NOTCH1-unmutated cases (p = 0.002) and the general group (p = 0.013). SNP rs3124591 did not affect the risk of CLL and survival parameters of the patients. At the same time, differences were found in the frequency of IGHV gene usage and in the structure of HCDR3 in carriers of individual genotypes. Conclusion: The frequency of NOTCH1 mutations in 3′UTR region was low. Our findings confirmed current data on the association between the structure of the B-cell receptor and the appearance of NOTCH1 mutations. Some features of HCDR3 structure were identified in carriers of TT and CC genotypes of rs3124591.
Submitted: May 29, 2018.
*Correspondence: E-mail: firstname.lastname@example.org
Abbreviations used: CLL — chronic lymphocytic leukemia; HCDR3 — heavy chain complementarity-determining region 3; ICN1 — intracellular fragment of NOTCH1; IGHV — immunoglobulin heavy-chain variable (genes); OS — overall survival; PFS — progression-free survival; TTT — time-to-treatment; UM — unmutated.
Deregulation of NOTCH1-signalling pathway is common in chronic lymphocytic leukemia (CLL). This gene encodes transmembrane receptor, and active intracellular fragment of NOTCH1 (ICN1) releasing after ligand binding, conformation changes and proteolytic cleavages translocates to the nucleus, where it mediates the transcription of target genes involved in cell differentiation, proliferation and apoptosis [1, 2]. Expression of ICN1 was found in 50.5% of CLL cases lacking NOTCH1 gene mutations regardless of mutation status of immunoglobulin heavy-chain variable (IGHV) genes . Prognostic significance of such constitutional ICN1 expression in CLL is not yet known.
Activated NOTCH1 mutations in CLL were first found through whole-genome sequencing by two independent groups in 2011 [4, 5]. They were commonly represented by a single 2-bp deletion (c.7544_7545delCT, P2514fs) in exon 34 and resulted in the removal of C-terminal PEST [proline (P), glutamic acid (E), serine (S), and threonine (T) rich] domain involved in proteasomal degradation of ICN1. NOTCH1 c.7544_7545delCT mutations occur in 8–11% in newly diagnosed CLL, 10–15% at the time of first treatment and 15–20% at chemorefractoriness and, according to the data of several authors, are associated with short progressive-free (PFS) and overall survival (OS) [6, 7]. Later, in the minority of CLL cases point mutations in the 3´UTR region of NOTCH1 were found that resulted in the removal of C-terminal PEST domain . We revealed c.7544_7545delCT mutations of NOTCH1 in 13.4% of untreated CLL patients. Associations between the presence of NOTCH1 mutations and unmutated (UM) IGHV genes, more advanced stage of the disease, higher initial WBC count, bulky disease, short time-to-treatment (TTT) period and PFS were found . The aims of present work were to evaluate the frequencies of mutations in the 3´UTR region of the NOTCH1 gene (9:136,495553–136,495994) in Ukrainian CLL patients, the distribution of rs3124591 genotypes located in that area, and association of NOTCH1 mutations with structure of B-cell receptor.
MATERIALS AND METHODS
Samples of 87 previously untreated CLL patients with UM IGHV genes and without mutations in hotspot regions of TP53, SF3B1, and NOTCH1 (exon 34) genes were selected for the study from the total group of 237 patients. Such choice of patients was based on the current data on a rare association of the NOTCH1 mutations with mutations of TP53 and SF3B1 genes, and their prevalence in cases with UM IGHV genes. All patients (69 males and 18 females) were referred to the National Research Center for Radiation Medicine, Kyiv, during the period of 2002–2016. The study was approved by the local Ethics Review Committee, and all patients gave informed consent prior to participation in it. The diagnosis of CLL was based on clinical history, lymphocyte morphology, and immunophenotypic criteria.
Genomic DNA for molecular analysis was extracted from peripheral blood mononuclear cells with the QIAamp Blood Mini Kit (Qiagen, Crawley, United Kingdom) according to the manufacturer’s protocol. Screening for presence of TP53, SF3B1 mutations and NOTCH1 mutations in exon 34 as well as mutational status of IGHV genes was performed in all 237 patients.
IGHV-D-J rearrangements were ampliﬁed according to the BIOMED-2 consortium rules  as described previously . IGHV rearrangements were analyzed by IMGT/V-QUEST . Mutational status of TP53 gene was performed for exons 3 to 10 as described previously . SF3B1 mutations were analyzed in exons 14, 15 and 16 by PCR ampliﬁcation followed by direct sequencing with the BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems) according to Rossi et al. . It should be noted that sequencing was carried out in those regions of TP53 and SF3B1 genes where the vast majority of mutations is located (> 95%) [15, 16].
NOTCH1 mutations were analyzed in the hotspot c.7282_7680 region in exon 34 of NOTCH1 gene by PCR amplification followed by direct sequencing as described previously . Detection of mutations in the 3´UTR region of NOTCH1 gene (9:136,495553–136,495994) was performed by direct sequencing according to Puente et al.  using original primers. PCR conditions and primers are described in: dx.doi.org/10.17504/protocols.io.qfhdtj6.
Analysis of rs3124591 genotypes was performed using SNPstats tool (bioinfo.iconcologia.net/snpstats/start.htm).
TTT period, PFS and OS were estimated by the method of Kaplan and Meier and assessed by the log-rank test.
The frequency of NOTCH1 mutations and association with HCDR3 structure. Mutations in the 3´UTR region of the NOTCH1 were revealed in three of 87 CLL patients (3.4%). Non-coding mutations were represented by 139390152T>C (two cases) and 139390145T>C (one case).
Two cases with non-coding mutations were related to subset #1 of stereotyped B-cell receptors, and one case belonged to stereotyped subset #28a (according to classifications of Stamatopoulos et al. , Murray et al. , Bomben et al. ).
To study the associations of NOTCH1 mutations and the stereotyped B-cell receptors, we have added to the studied group 30 UM IGHV cases with previously revealed c.7544_7545delCT mutations from the total group of 237 CLL patients. Thus, combined group of 117 UM-IGHV cases included 33 cases with NOTCH1 mutations and 84 cases without NOTCH1 mutations in exon 34 and in the 3´UTR region. It was divided into three subgroups according to the principle used by Sutton et al. .
The frequency of NOTCH1 mutations in subgroup of 22 CLL cases with UM IGHV genes of I phylogenetic clan (comprising IGHV1, IGHV5 and IGHV7 genes), except IGHV1-69, was the highest (13 of 22 cases; 59.1%). This subgroup included 9 cases of subset #1 (five with NOTCH1 mutations), 2 cases of subset #28a (both NOTCH1-mutated), one case of subsets #59, #95, #UA/ref4  (all NOTCH1-mutated), six cases showing homology with CLL sequences represented in the databases, but did not include in the stereotyped subsets (three with NOTCH1 mutations), and two NOTCH1-UM heterogeneous cases. Lengths of HCDR3 in NOTCH1-mutated cases varied from 12 a.a. to 21 a.a. and did not differ from NOTCH1-UM cases (15.66 ± 0.98 a.a. vs 18.55 ± 1.93 a.a.; p = 0.165).
Additional NOTCH1-mutated CLL case, not included in this subgroup, in its structure of IGHV-D-J rearrangement was close to cases of subsets #1 (structures of HCDR3 in CLL cases with NOTCH1 mutations are available by request).
In subgroup of 37 CLL cases with UM IGHV1-69 gene, NOTCH1 mutations were revealed in 12 cases (32.4%). NOTCH1-mutated cases belonged to subsets #6 (1 of 3), #9 (1 of 3), #UA/ref2 (1 of 2), #UA/ref11 (one case). Additional one NOTCH1-mutated case was revealed among six cases homologous with CLL sequences represented in the databases, but did not include in the stereotyped subtypes, and seven NOTCH1 mutations were found among fourteen CLL cases with heterogeneous HCDR3s. Without NOTCH1 mutations there were cases of subsets #7 (n = 5), #UA4 (n = 1), #UA6 (n = 1), and #UA7 (n = 1). Lengths of HCDR3 in NOTCH1-mutated cases varied from 17 a.a. to 25 a.a. and did not differ from NOTCH1-UM cases (21.08 ± 0.75 a.a. vs 21.64 ± 0.71 a.a.; p = 0.635).
In subgroup of 57 CLL cases with other UM genes (IGHV3, n = 39; IGHV4, n = 16; IGHV2, n = 2) NOTCH1 mutations were present in seven cases (12.3%). Two NOTCH1-mutated cases belonged to subsets #22 (1 of 3) and #25 (one case), two had homology with CLL sequences represented in the databases, but not included in the stereotyped subtypes (total n = 15), the three remaining cases had heterogeneous HCDR3s (total n = 25). CLL cases from subsets #7 (n = 2), #9 (n = 1), #31 (n = 1), #41 (n = 2), #44 (n = 1), #50 (n = 1), #109 (n = 1), #UA5 (n = 2), and #UA9 (n = 2) were NOTCH1-UM. Mean lengths of HCDR3 did not differ in NOTCH1-mutated (21.14 ± 0.98 a.a.) and NOTCH1-UM cases (20.60 ± 0.64 a.a.; p = 0.760).
Differences in the frequency of NOTCH1 mutations in subgroups of cases expressed UM IGHV genes of I clan (except IGHV1-69), IGHV1-69 gene and other UM IGHV genes were significant (p = 0.001). The distribution of UM IGHV genes of the selected subgroups among NOTCH1-mutated cases was significantly different comparing with the distribution in NOTCH1-UM cases, in the analyzed group of 117 CLL cases as well as in total group of 237 CLL cases (Table 1).
Table 1. The distribution of UM IGHV genes in studied CLL patients
Analysis of rs3124591 SNP distribution and its association with survival parameters and IGHV gene usage. The distribution of rs3124591 genotypes was as follows: CC genotype — 20 cases (21.5%), CT genotype — 41 cases (44.1%), and TT genotype — 32 cases (34.4%) and did not deviate from the Hardy — Weinberg equilibrium (p = 0.04). In comparison with healthy individuals of European ancestry retrieved from the 1000 Genomes Project dataset (CC genotype 26.2%; CT genotype 47.4%; TT genotype 26.0%; www.1000genomes.org) no significant differences were found. All three cases with non-coding NOTCH1 mutations were represented in carriers of TT genotype.
The impact of rs3124591 on duration of TTT period, PFS and OS was insignificant (p = 0.425, p = 0.380, and p = 0.722, correspondingly).
The spectrum of used IGHV genes tended to be narrower in CC homozygotes than in carriers of CT and TT genotypes (correspondingly 22, 20, and 8 IGHV genes, p = 0.125; p = 0.043 in comparison TT vs CC genotype carriers) (Table 2).
Table 2. The distribution of UM IGHV genes in carriers of different rs3124591 genotypes
Expression of only two IGHV genes was detected in more than half of CC homozygotes (8 cases with IGHV1-69 and 4 cases with IGHV3-11 gene).We found a reduced IGHV1-69 gene usage in carriers of TT genotype compared to carriers of CT and CC genotypes (12.9%, 35.1%, and 42.1%, correspondingly; p = 0.038). The distributions of IGHD and IGHJ genes were comparable in carriers of different genotypes. It should be noted that the frequencies of IGHV and IGHJ genes usage and different stereotyped subsets in this studied group did not differ from those in the previously studied large group cohort.
The HCDR3 length did not differ in carriers of different genotypes with UM IGHV genes (CC genotype 21.43 ± 1.14 a.a.; CT genotype 21.05 ± 0.73 a.a.; TT genotype 19.16 ± 0.76 a.a.; p = 0.140). However, the number of N nucleotides inserted in the VHD junctions was significantly less in carriers of TT genotype than in carriers of CT and CC genotypes (Table 3).
Table 3. N-nucleotide additions and exonuclease activities in the VHD and DJH junctions in carriers of different rs3124591 genotypes
The comparison of CLL sequences with non-CLL sequences available from public databases showed that most cases that had HCDR3 homology with antibacterial or antiviral Ig clones were present in TT homozygotes (22.6% vs 2.7% in carriers of CT genotype and 0% in carriers of CC genotype, p = 0.006; Table 4). All CLL sequences homologous with autoreactive clones were revealed in carriers of TT (9.7%) and CT (10.8%) genotypes. A number of CLL cases that were similar to Ig sequences expressed by normal B-cells (elderly, neonate, cord blood, tonsils) and to Ig sequences from patients with X-linked hyperIgM did not differ in carriers of different rs3124591 genotypes (Table 5, 6).
Table 4. HCDR3 homology in carriers of different rs3124591 genotypes
Table 5. Structure of HCDR3 region in CLL cases with NOTCH1 mutations (MAS — mean alignment score)
Table 6. CLL cases showing HCDR3 homology with immunoglobulins of known specificity (MAS — mean alignment score)
To investigate non-coding mutations in the 3´UTR region of the NOTCH1, previously untreated CLL cases with no mutations in TP53, SF3B1, and NOTCH1 (exon 34) genes were selected. All of them expressed UM IGHV genes. Such choice was due to the literature data on a rare association of the NOTCH1 mutations with mutations of the other genes. In addition, NOTCH1 mutations are mainly detected in CLL patients with UM IGHV genes. For example, in the study of Rossi et al. 76.5% of CLL cases with NOTCH1 mutations in exon 34 expressed UM IGHV genes, and more than 90% of them did not have mutations of TP53 gene . In the study of Schnaiter et al., in none of 97 fludarabine-refractory CLL patients concurrent NOTCH1 and SF3B1 were found, but 23.1% of NOTCH1-mutated cases had simultaneously TP53 gene mutations. All revealed NOTCH1 mutations in this study were in patients with UM IGHV genes . In our group of 237 CLL patients, mutations of TP53, SF3B1, and c.7544_7545delCT of NOTCH1 were found in 12.1%, 10.8%, and 13.1% of cases, correspondingly. Only single case harbored mutations in both NOTCH1 and TP53 genes and two cases — in NOTCH1 and SF3B1 gene. All except one NOTCH1-mutated cases had UM IGHV genes. Therefore, we investigated a group of CLL patients with an increased chance to identify mutations in the 3´UTR region of NOTCH1 gene.
Non-coding mutations in the 3´UTR region of NOTCH1 gene were first identified by Puente et al. . Authors revealed 12 NOTCH1 mutated cases of 176 CLL patients with UM IGHV genes (6.7%). The most frequent recurrent non-coding mutation was 139390152T>C, and two cases had 139390145 and 139390143 point mutations. Nadeu et al. detected 3´UTR mutations in 22 of 391 CLL patients (5.6%) . We have identified two cases with 139390152T>C mutation and one case with 139390145T>C mutation among 87 IGHV-UM CLL cases (3.4%). The frequency of non-coding mutations in our group and their localization were comparable with previously published data.
Our data confirmed current data on the association between the structure of the B-cell receptor and appearance of NOTCH1 mutations. In one of the largest series of studied stereotyped subset CLL cases (565 cases assigned to one of 10 major stereotyped subsets) the skewed distribution of NOTCH1 mutations within exon 34 was found. Enriched mutations were cases belonged to subsets #1, #59, #99, #6, and #8. NOTCH1 mutations were relatively infrequent in cases from subsets #3, #5, #7 and, especially, from subsets #2 and #4 . In our group, five of the nine cases from stereotyped subset #1 were NOTCH1-mutated: three had mutations within exon 34 and two had non-coding mutations in the 3´UTR region. On the contrary, NOTCH1 mutations were not found among nine cases of subset #7. The number of cases belonged to other subsets was too low for comparison. However, the observed differences in the frequency of NOTCH1 mutations in subgroups of cases that expressed UM IGHV genes of I clan (except IGHV1-69), UM IGHV1-69 gene and other UM IGHV genes (59.1%, 32.4%, 12.3%, correspondingly) cannot be explained solely by the presence in the first subgroup of the cases belonged to subset #1. Excluding cases of subset #1, NOTCH1 mutations were detected in eight of 17 remaining cases of the first subgroup (47.0%). Both cases of subset #28a were NOTCH1-mutated (mutations within exon 34 and non-coding region). The occurrence of UM IGHV genes of I clan (except IGHV1-69) among NOTCH1-mutated cases was twice as high as in the general group. Therefore, it can be suggested that expression of UM IGHV genes of I clan (except IGHV1-69) is a risk factor for the presence of NOTCH1 mutations. On the contrary, the probability of detecting NOTCH1 mutations was lower in the carriers of other UM genes (it should be noted that subset #8 was absent in our cohort).
When searching for non-coding mutations in the 3´UTR region of the NOTCH1, we paid attention to rs3124591, which is localized in the amplified region. Preliminary data suggested its functional significance. Quan et al. found association of CC genotype of rs3124591 (16.4% among cases and 11.1% in the control group) with the risk of lung cancer in northeast Chinese non-smoking females . In study of Gao et al., also performed in Chinese population, CC genotype of rs3124591 was absent in the control group (n = 100) and among patients with invasive ductal carcinoma (IDC, n = 100) and ductal carcinoma in situ (DCIS, n = 50) . The C allele was significantly associated with high risk of DCIS, but not IDC; the TC genotype was significantly associated with an increased risk of IDC and DCIS and poorly differentiated IDC. In addition, Notch1 protein expression was significantly higher in DCIS patients with the TC genotype. Although Notch1 protein expression was higher in IDC patients with the TC genotype, this association did not reach significance. Authors concluded that the impact of the rs3124591 variant on Notch1 protein expression mainly occurs early in IDC development.
In our study, rs3124591 did not affect the risk of CLL and survival parameters of patients. At the same time, differences were found in the frequency of IGHV gene usage and in the structure of HCDR3 in carriers of individual genotypes. Leukemic cells of CC homozygotes expressed the most limited spectrum of UM IGHV genes (mostly IGHV1-69 and IGHV3-11), and their HCDR3 sequences were homologous only with the sequences of normal B-cells. Conversely, leukemic cells of TT homozygotes used the largest number of UM IGHV genes, more frequently had HCDR3 homology with antibacterial or antiviral Ig clones, had less N nucleotide additions in DJH junctions, and the number of sequences that lacked N additions at DJH junctions had a tendency to be higher. By two last parameters, they were a bit like a memory B cells, characterized by Tian et al. . It is known that Notch1 signaling regulates B and T lymphocyte development [27–30]. Taking into account data of Cao et al.  on the functional significance of rs3124591, we hypothesized that rs3124591 could influence on the selection of B-cell clones during early stages of CLL development. This assumption are in an agreement with reports regarding the role and possible mechanisms of NOTCH signaling in regulation of the normal B-cell repertoire, summarized by Cruickshank and Ulgiati .
The authors are grateful to Mr. Thomas Harms, President of Charitable Organization “KIHEW-Kinderhilfe” (Germany) who provided support by reagents for fulfillment of the work.
No Comments » Add comments