Methyltransferases are a critical class of highly evolutionarily conserved enzymes capable of exerting a wide range of regulatory effects through their ability to transfer a methyl group onto a substrate. The human genome is estimated to contain 208 genes encoding methyltransferases (1) where the majority of these contain the canonical Rossmann-fold S-adenosyl-L-methionine (SAM) binding domain (2, 3). The activity of methyltransferases has been demonstrated against a broad range of substrates including nucleic acids (DNA/RNA), proteins, lipids, and small molecules (4–6). Arguably the most researched form of methylation is performed by DNA methyltransferases. Both the enzymes that methylate the DNA and the proteins that recognize this modification play critical roles in normal development through their ability to modulate transcriptional activity (e.g., gene silencing through the hypermethylation of gene promoter regions) (5). The methylation of protein substrates, notably histone tails, has also been investigated extensively and been shown to play an equally important role to DNA methylation in maintaining epigenetic regulation of genomes through a complex network of enzymatic “writers” and “readers”. Together, DNA and histone methylation provide a highly integrated system to functionalize genomes by controlling accessibility of the DNA for transcription that in turn influence cell fate and development through coordinated gene expression programs. More recently, another layer of methylation control over the transcriptome has been uncovered, namely RNA methylation. These base modifications, many of which are still being characterized, have been shown to have an equally broad range of impacts on mRNAs including their splicing, nuclear export, translation, and degradation (7). Given their diverse range of normal activities, it is perhaps not surprising that methyltransferase genes have been reported to have potential roles in many different cancers when dysregulated or mutated (4, 8, 9). In the context of malignancies, the identification of the substrate specificity of each enzyme is obviously of great interest, however current predictive methods are insufficiently accurate for this task for several reasons (3, 8). For instance, generating accurate protein structures where multiple subunits are involved in order to predict substrates remains challenging. In addition, the large degree of sequence variation in methyltransferase domains that target the same substrate means that there is a limited degree of certainty with respect to predicted substrates. As a result, the mechanistic activity of many methyltransferase enzymes in cancers remains unknown.

Within the SAM-dependent methyltransferases, one sub-group of enzymes whose importance in cancer has recently been elucidated are those encoded by the methyltransferase-like (METTL) genes. This family of enzymes contains 33 members which share little sequence conservation at the protein level outside of their SAM binding domain (8, 10, 11). The best-characterized members of the METTL family are METTL3 and METTL14, which form a heterodimer capable of catalyzing the N⁶-methyladenosine modification (m⁶A) on adenosine residues in RNA molecules (12). This modification can in turn be recognized by a wide range of modification “readers” that can influence transcript splicing, transport, stability, and translation (6). Other members of the METTL family have subsequently been shown to catalyze not only other nucleotide modifications such as N⁴-methylcytidine (METTL15) or O²-methyluracil (METTL19) but to also methylate protein substrates on different residues (METTL10, METTL11A/B, METTL20) (8, 11). The substrate and activity of other family members however, including Thiol Methyltransferase 1B (TMT1B, also known as Associated with Lipid Droplets 1; ALDI, and Methyltransferase-Like 7B; METTL7B), the focus of this review, remain poorly defined.

In the human genome, the TMT1B gene encodes a 244 amino acid protein that is most closely related to its putative paralog TMT1A, with the encoded proteins sharing 59% residue identity (13). However, despite the identical size of both proteins, and similar predicted localization to the endoplasmic reticulum (ER), the fact that TMT1B and TMT1A show greater similarity to their mouse orthologs (84% and 85% respectively) than to each other argue their functions have likely diverged over time. This argument is supported both by the biochemical studies discussed later and the divergent expression of both genes in normal tissue. While TMT1A is highly and ubiquitously expressed, TMT1B’s expression is largely restricted to the heart, liver, and colon (14). Given these differences, it is not possible to infer the function of TMT1B from its closest gene family relatives.

TMT1B was first identified in 2004 and characterized as a putative methyltransferase localized to the perinuclear face of the Golgi body in rat kidney cells (15). Later functional studies with purified proteins confirmed that TMT1B’s methyltransferase activity is SAM-dependent (13, 16, 17), while microscopy studies with cultured cells refined its subcellular localisation to the perinuclear face of the ER (13, 17, 18), and the surface of lipid droplets (LD) (18, 19), rather than the Golgi body (18, 20). Differential expression of TMT1B has been found to occur within several cancer types (4), including gliomas (21–25), non-small cell lung cancer (16, 23, 26–29), and acute myeloid leukemia (30, 31). TMT1B has been proposed as a biomarker for many of these cancers since its increased transcription and protein abundance is frequently correlated with an advanced TNM (Tumour, Node, Metastasis) stage of tumour development and poor patient overall survival (21–27, 29, 31–36), risk of recurrence (35) and drug resistance (16, 37). Despite these characterizations, the exact substrate of TMT1B, and its role in the various cancers where it is highly expressed, have not yet been defined. In this review we summarize the findings of several functional and biochemical studies of TMT1B to provide a comprehensive view of its activity and potential insight into its role in different cancers.

Given the in vitro biochemical evidence for different potential substrates for TMT1B, the question arises whether any of the functional studies of TMT1B might provide further insight into its substrate, and therefore biological role. While more detailed studies would be required to precisely define its role, TMT1B activity could potentially touch on multiple cellular functions recognized as “hallmarks of cancer” such as involvement in proliferation, migration and invasion, the tumour immune microenvironment (22, 24), as well as cellular metabolism (13, 16). Because of its overexpression in a wide range of cancers, most of the published functional data is in this context and is broadly summarized in Table 1 and briefly discussed here.

A further avenue through which to try and gain insight into the activity of TMT1B is through its sub-cellular localization. TMT1B was originally identified as a constituent of the Golgi complex and while later studies have linked its expression to the maintenance of Golgi integrity (20) this same study in HeLa cells found that TMT1B (and TMT1A) was localized to the ER, and not the Golgi. This discrepancy led the authors to speculate that the target of TMT1B might be a protein which shuttles between the Golgi and ER. It has been suggested that the largely hydrophobic N-terminus of TMT1B might play a role in directing its localization specifically to LDs. This view is supported by experiments with a GFP-tagged version of TMT1B expressed in COS cells where the protein was either in the membrane of the ER when cells were serum starved or in LDs when cells were pre-treated with fatty acids (FA) (18). As noted earlier, similar observations were made in neuronal cells (17), where TMT1B also translocated from the ER to LDs under high lipid conditions. LDs are thought to function as transport or communication hubs between organelles, including mitochondria (52, 53), through their transport of associated proteins (54). It is therefore possible to speculate that in the case of TMT1B, LDs might provide a mechanism for the enzyme to move to different subcellular locations. In this context, it is interesting that a mass-spectrometry based study looking at interacting proteins of METTL family members in HeLa cells only detected robust interaction between TMT1B and TMEM126A, an uncharacterized transmembrane mitochondrial protein (10). In the context of cancerous cells, where TMT1B often shows increased expression, lipid metabolism shows a definitive shift towards de novo FA synthesis—leading to a higher number of LDs (55)—as compared to exogenous accumulation of FAs in non-tumour cells in order to compensate for its higher energy consumption and increased cell division (56). Together, these observations suggest that the subcellular movement of TMT1B could potentially be linked to its possibly diverse substrates. At the same time, very few studies have conducted detailed studies of the subcellular movement of TMT1B, and none have been performed in the context of normal versus cancerous cells to elucidate whether its overexpression might lead to an aberrant localization and function specific to tumours.

Despite the relatively sparse information on its biochemical substrates, functional studies characterizing the impact of the loss of expression of TMT1B have demonstrated its ability to impact multiple oncogenic pathways and play a role in cancer progression (8, 20). Overall, TMT1B expression in cancer is tied to pro-proliferation pathways, pro-migratory and pro-invasion functions, and finally to ROS scavenging and H₂S metabolism. The question is therefore raised about whether TMT1B directly affects all these pathways (either alone or in a complex) or which of these effects may be indirect consequences. The answer to this question is likely linked to both its location in the cell as well as its interactions, whether with alkyl thiol molecules or proteins, in these subcellular locations.

There are at least two potential scenarios that could integrate the published data on TMT1B activity in cells and the phenotypes of its loss (see Figure 1 ). In the first case, its primary activity may be through the regulation of H₂S metabolism, the direct methylation of which has been demonstrated. Through the maintenance of a balanced H₂S environment, TMT1B could indirectly modulate ROS (42) and lipid (57) levels thus allowing for indirect regulation of cellular growth, migration and invasion (56, 58, 59). TMT1B’s role in H₂S metabolism would provide an advantage to cancer cells when upregulated and could explain the protein interaction with mitochondrial membrane protein TMEM126A (10). The second possibility by which TMT1B can have such a range of impacts would be through its role as a protein methyltransferase. In this case, TMT1B could be shuttled between the Golgi complex, the ER and LDs (as previously shown) depending on the metabolic state of the cell, allowing for it to methylate a wide range of proteins depending on its current localization. These proteins could then, in turn, affect pathways involved in proliferation, migration and invasion, and ROS metabolism. It is important to note that these two scenarios are not mutually exclusive but rather TMT1B may methylate both thiol and protein substrates depending on circumstance. Given the divergent in vitro data published on its activity, additional in vivo targeted studies would be required to confirm published in vitro observations.

Regardless of its direct activity, it is clear that patients with TMT1B-high cancers generally show significant decreases in overall and disease-free survival (22–24). Additionally, regardless of if this is a direct cause or simply an effect of TMT1B expression, tyrosine kinase inhibitor resistance in lung adenocarcinoma patients was identified and the authors concluded that the resistance was due to TMT1B’s increased expression (16). The authors suggest that this might be due to TMT1B’s ties to an antioxidative pathway, although independent of its canonical regulator NRF2, since sensitivity to TKIs was reversed when cells were treated with ROS-scavenging inhibitors or TMT1B expression was ablated (16). This study suggests that from a therapeutic viewpoint of TMT1B over-expressing cancers, combinatory therapies would likely be required to block targets downstream of aberrant growth signals (e.g., tyrosine kinase receptor activation) such as ROS scavenging targets.

In conclusion, the identification of TMT1B as a biomarker in an increasing number of poor prognostic cancers recognizes this methyltransferase to be of importance in cancer progression. Continued comprehensive characterization of TMT1B’s activity, and the identification of its methylation targets will provide a better understanding of how this methyltransferase can impact such diverse and key pathways involved in cancers, and potentially provide mechanisms of inhibition for TMT1B-high cancers.

SD: conception of study, literature search, data interpretation, writing of initial draft. BW: critical revision and editing of draft, preparation of figure, supervision of project. All authors contributed to the article and approved the submitted version.