Molecular genetic aberrations in the pathogenesis of multiple myeloma

The co-occurrence of two or three high-risk genetic alterations is defined as double-/triple-hit MM. Based on the criteria stated in the IMWG, clinical outcomes of double-hit MM are similar regardless of whether they are classified as high-risk or low/standard-risk MM. Thus, defining the molecular features of double-hit MM is important to improve risk stratification and outcome prediction in MM. Double-hit MM is found in 1 of 33 newly diagnosed multiple myeloma (NDMM) [40]. According to Baysal et al. [52], the OS of MM patients with double-hit, single high-risk, and no high-risk MM were 6 months, 32 months, and 57 months, respectively. The same study found that the hazard ratio was worsened in triple-hit (7.30) than double-hit (5.55) and single high-risk (1.42) MM patients [52]. Another study revealed a close relationship between the co-occurrence of translocation t(4;14) and TP53 mutation or del(17p) and TP53 mutation, where they were found to confer poor prognosis and OS in the patients [53]. By using genome-wide NGS analysis, two genomic variables in double-hit MM were defined according to the background of International Staging System (ISS) III [54]. They were the biallelic inactivation of TP53 and amplification (≥4 copies) of CKS1B (1q21). Biallelic inactivation of TP53 has an adverse prognosis compared to wild-type or monoallelic inactivation of TP53 [54]. Amplification of CKS1B has a shorter progression-free survival (PFS) and OS than a gain of 1q. There was no significant change in PFS when amplification of 1q co-occurred with biallelic TP53, del(17p), t(4;14), or t(14;16) [54]. Double/triple hits exhibited extra-high-risk features in MM, and the higher is the number of “hits,” the lower is the OS of the patients. In addition to the number of “hits,” the type of double/triple hits also significantly influences the prognosis outcomes and OS of the patients. Thus, patients presented with double/triple hits may require more intensive treatment than those with standard-risk MM, suggesting the importance of defining subtle genetic differences in patients using high-end technology to improve prognostic and treatment outcomes.

Transcriptome changes play a critical role in myelomagenesis as early as in the monoclonal gammopathy of undetermined significance (MGUS) stage [55]. In the past decades, GEP or microarray has emerged as one of the most popular approaches in the identification of transcriptome changes, molecular mechanisms, and pathways underlying human cancers. GEP has contributed to the characterization of disease subtypes, risk stratification, prognosis, and outcomes in MM patients. One of the most prominent results from GEP was the discovery of a 70-gene signature model by the University of Arkansas for Medical Sciences (UAMS) in 2007 [56]. The 70-gene signature was found to be superior in risk and prognostic stratification of NDMM. They found that most of the crucial genes related to disease progression were localized in chromosome 1. From the same study, 17-gene subsets for high-risk MM were also revealed. Among the important genes found in the 17-gene subset prediction model are CKS1B, ASPM, and CTBS.

Another significant study was reported by Broyl et al. [57]. They discovered seven unique clusters, which were associated with NDMM based on the gene expression profiles: i. translocation [MMSET/t(4;14), MAF/t(14;16)/t(14;20), cyclin D1/t(11;14), and cyclin D2/t(6;14)]; ii. hyperdiploidy (HY); iii. proliferation-associated genes (PR); iv. low percentage of bone disease (LB); v. overexpression of cancer testis antigens (CTAs); vi. NFκB pathway; and vii. overexpression of protein tyrosine phosphatases (PRLs) [57]. Among the seven clusters, three were newly identified (v–vii), while the rest were consistent with the UAMS classification [58].

The EMC-92 gene signature prediction model for high-risk MM was developed by another group of scientists from the Netherlands based on the findings from the GEP [59]. Survivin/BIRC5, FGFR3, and STAT1 are the key genes in the EMC-92 model. Recent combination analysis of EMC-92, UAMS70, ISS, and FISH [t(4;14) & del(17p13)] has shown that EMC-92 gene classifier + ISS was the most superior marker for risk stratification in MM patients [59, 60]. This sheds light on the application of GEP in prognosis and risk prediction, replacing FISH, which is relatively expensive and laborious. FISH, however, can be used as a confirmation test if required.

Although GEP has been successfully used to characterize MM into different clinicopathological subtypes and molecular signatures for diagnosis, staging, risk stratification, and prognostication of MM, data generated by GEP are inadequate to fully delineate the molecular biology of this highly genetically complex disease. In recent years, NGS has become a promising tool in the study of IgH translocations, V(D)J clonal rearrangements, IgH isotype, CNAs, and somatic mutations simultaneously in a more refined manner.

The very first mutational profiles generated by NGS data in the past 10 years revealed that there is no specific or unique mutation in MM. The observation remains valid in the latest findings [61, 62]. To date, approximately 250 mutated genes have been found in MM, and about a quarter of them are identified as driver genes [62]. N-Ras, K-Ras, DIS3, FAM46C, TP53, and BRAF are considered the top recurrent or driver genes in MM, with a mutation frequency range between 6%–25% each (Figure 2) [53, 61, 63, 64]. Mutation in other driver genes such as TRAF3, SP140, IRF4, JAK3, and PPGFRA are rare with low recurrence rates (<5%) [65, 66]. Other newly identified potential driver genes are IDH1, IDH2, HUWE1, KLHL6, and PTPN11 [67]. N-Ras, K-Ras, and BRAF are key components of the MAPK pathway [61, 68, 69]. Studies showed that the MEK-ERK signaling pathway was affected by BRAF and RAS mutations in >50% of the cases [70]. The most common site of N-Ras mutation is in codon Q61, while in K-Ras, it is in codons G12, G13, and Q61 [67]. On the other hand, BRAF mutation is predominantly seen in codon D594 (92%) in the t(14;16) group and BRAF V600E in the t(4;14) group [67]. The discovery of mutations in BRAF allowed many MM patients to benefit from BRAF inhibitor (vemurafenib) treatment [10].

Figure 2.

NGS findings not only confirm the heterogeneous complexity of MM but also show that most of the driver mutations are present in the subclonal population, and multiple mutations are detected in different genes within the same pathway, suggesting a diverse pattern of clonal evolution in MM, which evolves through space and time [10, 70]. The main clonal evolution in MM includes neutral evolution, branching evolution, and linear evolution [10, 71]. Neutral evolution occurs when all descendant subclones show a similar ability to survive under certain circumstances, and the evolutionary changes are not caused by selective pressure. Branching evolution is the early divergence of subclones with different mutations, which evolves further over time, possibly driven by the selective pressure of the bone marrow microenvironment and treatment, or inherent tumor characteristics, or both. In linear evolution, a single subclonal population is fully substituted by another highly adapted subclone [10, 71]. The main clonal evolution models in MM are visualized in Figure 3. Interestingly, recent NGS studies have shown that the emergence of multiple selective clonal sweeps derived from single cells can drive relapse even after 10 years of remission, suggesting the dormancy of resistant clones [72].

Figure 3.

Despite the discovery of driver gene mutations and clonal evolution in MM, another major contribution by NGS was the uncovering of complex SVs in MM. Complex SVs are usually difficult to detect using conventional sequencing approaches. Common complex SVs in MM include chromothripsis, chromoplexy, and multiple templated insertions [73]. Generally, MM patients present with at least one complex SV (80%) [38]. Chromothripsis, a form of chromoanagenesis, was first detected in chronic lymphoblastic leukemia (CLL) [74]. Chromothripsis is characterized by ten to hundreds of chromosomes breaking and rejoining in confined genomic regions in one or a few chromosomes (Figure 4). In contrast, chromoplexy involved a larger number of chromosomal rearrangements than chromothripsis (Figure 4). It was first identified in prostate cancer using whole-genome sequencing (WGS) [75]. Chromothripsis and templated insertions were found to be associated with clonal events, which occur early in the disease course, while chromoplexy is a late event that is possibly related to disease progression, drug resistance, and relapse [38]. There is a strong association between chromothripsis and adverse clinical outcomes in NDMM [76]. Templated insertions are less common than chromothripsis (19% vs 24%) [77]. It is characterized by focal gains bounded by translocations, resulting in the concatenation of amplified segments from two or more chromosomes into a continuous stretch of DNA, which is inserted back into any of the related chromosomes. SV hot spots frequently implicate known MM oncogenes or tumor suppressor genes such as Ig, MYC, CCND1, MMSET, IRF4, MAP3K14, FAM46C, CDKN2C, CYLD, and SP140 [38, 78]. Recently, novel SVs involving critical genes in immune-based therapies TNFRSF17, SLAMF7, and BCMA were also revealed [78]. Patients who harbored SVs involving these genes tend not to respond to elotuzumab, BiTEs, or CAR-T cell therapy, or to their combination. NGS of MM also discovered various SVs hot spots in novel potential driver genes such as IRF2, BTG2, SOX30, NEDD9, KLF13, and TNFRSF13B [78].

Figure 4.

NGS also contributes to the revelation of the activity-induced deaminase (AID)/apolipoprotein B mRNA editing enzyme and catalytic polypeptide-like (APOBEC) mutational signatures in MM. AID or APOBEC plays an important role in mediating somatic mutations and genomic instability in MM. The role of AID in initiating oncogenesis of MM has long been studied. AID is an essential regulator of somatic hypermutation and class-switch recombination of Ig genes. AID initiates C to U mismatches in single-stranded DNA and preferentially deaminates WRC/GYW hot spots (W=A/T, R=A/G, Y=C/T) [79]. AID is also a crucial mutator for IgH/MYC chromosomal translocation and oncogene mutation in MM [80]. So far, many AID-target driver genes have been identified. Among the most significant were EGR1, IRF4, LTB, PIM1, TET2, and XBP1 [81]. Mutations in AID-target driver genes provide a selective growth advantage in the early phase of the disease [53].

On the other hand, APOBEC family proteins consist of 10 enzymes, namely, A1, A2, A3A, A3B, A3C, A3D, A3F, A3G, A3H, and A4. APOBEC is critical in mRNA editing and antiviral defense in the cells [82]. Dysregulation of APOBEC causes C>T, C>G, and C>A mutations in a TpC context and thereby resulting in copy number and mutational changes. In contrast to AID, APOBEC mutational activity is likely correlated with late events in disease progression [53]. High expression of APOBEC promotes DNA damage and increases mutational burden in MM (e.g., APOBEC3G) [82, 83]. Studies have shown that aberrant APOBEC expression is associated with translocations t(14;16)/MAF and t(14;20)/MAFB and a poor prognosis impact in MM [83]. However, only a small fraction of patients who presented with MAF/MAFB/MAFA translocations were related to high APOBEC activity (~23%), suggesting that other factors may be involved in regulating MAF-related translocations [84]. Co-occurrence of APOBEC activity and ISS stage III reduced OS in MM patients when compared to patients without the risk factors (53.8% vs 93.3%) [84]. Interestingly, APOBEC also triggers kataegis, a pattern of localized hypermutations characterized by C>T transitions at TpC dinucleotides. Kataegis has been observed at the MYC/IgK or IgL translocation breakpoints, implicating the co-occurrence of translocations and APOBEC-associated mutations in MM [85]. Suppression of APOBEC in MM cell lines was shown to reduce genomic and mutational changes by 30–65% [82]. A recent study that investigated the genomic spectrum of smoldering multiple myeloma (SMM) has shown that SMM patients with APOBEC-associated mutations demonstrated a shorter time of progression to MM [42]. In other words, the presence of APOBEC mutational signatures could be an indicator for higher risk of progression from SMM to symptomatic MM. Therefore, APOBEC inhibitors are potential therapeutic agents that can be used to suppress genome evolution in SMM or MM as a measure to prevent or delay the progression of the disease. However, this approach is still being developed. Nonetheless, due to the negative impact of APOBEC activity in prognostic and OS of MM patients, it is worth to include APOBEC analysis early on at the diagnosis stage to improve the treatment choice and efficacy in patients with high APOBEC activity.

The genomic spectrum of MGUS and SMM has been studied using NGS as well, although they are not as much as in MM. MGUS is a benign noncancerous condition, with a ~1% progression rate to MM per year. In contrast, SMM is a precancerous or early-stage MM, with a 10% annual progression risk to active MM in the first 5 years of diagnosis [86].

By comparing NGS data from paired SMM and MM from the same individual, Bolli et al. [53] postulated that most of the driver mutations such as hyperdiploidy, IgH translocations, 1q gain, and del(13q) in SMM were clonal and retained in MM. They found 2 unique clusters during the progression of SMM into MM: the static progression model, in which the same subclonal architecture in SMM was retained in MM; spontaneous evolution cluster, where the additional mutation in subclonal composition was acquired during disease progression [53]. Their findings support the clonal evolution patterns that have been described before [10]. Importantly, their results suggest different management and treatment strategies for the two clusters. Another study was published in 2020, which described the genomic signatures of SMM and its association with risk progression to MM [42]. The group analyzed the genomic profiles of 214 SMM patients using whole-exome sequencing (WES) or deep-targeted sequencing approaches. Their study findings were consistent with Bolli et al. [53], which demonstrated that most of the driver mutations previously defined in MM were already present in SMM patients. These driver mutations were also identified as independent risk factors of progression to MM [42]. In addition, their findings have shown that SMM, which presented with genetic alteration related to the MAPK pathway, DNA repair pathway, or MYC, has a higher risk of progression to MM.

In addition to comparing genomic data between SMM and MM, WGS has also been used to study the differences in the genomic landscape and temporal acquisition of myeloma-associated genomic events between clinically stable and progressive myeloma precursor conditions (MGUS and SMM) [87]. Compared with the clinically progressive myeloma precursor condition, the clinically stable myeloma precursor condition was associated with late initiation of the first clonal copy number alteration in patients’ life and absence or lower number of mutations in driver genes (genes involved in MAPK and NF-κB pathways, TP53), aneuploidy (gain of 1q, deletion of 6q, gain of 8q24, deletion of 8p, and deletion of 16q), templated insertions, chromothripsis, and APOBEC-associated mutational activity.