Multiple Myeloma (MM) is a hematological malignancy characterized by the accumulation of abnormal monoclonal plasma cells in the bone marrow (BM). It is the second most frequent hematological cancer and comprises 10% of all hematological malignancies, with defined clinical characteristics including hypercalcemia, renal failure, anemia, and bone lesions (CRAB) [1, 2]. Worldwide, an estimated 160,000 people were diagnosed with MM in 2020 [3].

The discovery of novel drugs, including proteasome inhibitors (PI; Bortezomib, Carfilzomib, and Ixazomib) and immunomodulatory drugs (IMiD; Thalidomide, Lenalidomide and Pomalidomide), has significantly altered the therapeutic landscape for MM in both the frontline and relapsed/refractory setting during the past two decades. The combined application of these drugs, together with the use of myeloablative chemotherapy and autologous stem cell transplantation (ASCT), has translated into prolonged overall survival (OS) rates with reduced toxicity and improved quality of life [4, 5]. More recently, immunotherapy has emerged as a powerful new tool to obtain durable responses in MM. This type of therapy includes monoclonal antibodies, immune checkpoint inhibitors, bispecific antibodies, chimeric antigen receptor T (CAR-T) cells, and peptide vaccines [6,7,8,9]. However, despite these new advancements, MM remains largely incurable due to either the occurrence of immune suppression or the development of drug resistance to multiple drug classes. With modern therapy, the first relapse typically occurs after about 3–4 years following initial diagnosis [2].

The (hypoxic) BM environment wherein the MM cells grow provides support and protection against different types of drugs. It consists of several cell types including BM stromal cells, endothelial cells, osteoclasts and osteoblasts. All these different cell types contribute to the growth and expansion of the MM clone, by providing nutrients and growth factors such as metabolites, amino acids, and cytokines. The main growth factors for MM cells include interleukin-6 (IL-6), insulin-like growth factor-1 (IGF-1) and vascular endothelial growth factor (VEGF). These growth factors will activate different signaling cascades with the ultimate goal to stimulate biogenesis and cell division [10].

Maintaining a stable proteome is essential for the growth and survival of every cell, yet protein synthesis (mRNA translation) and folding processes are inherently error-prone. The key steps in protein synthesis include initiation, elongation, termination and ribosome recycling [11]. Excessive protein synthesis has been associated with human cancers with elevated global translation, such as MM where there is a high production of immunoglobulins. The mammalian target of rapamycin (mTOR) kinase controls several factors involved in protein synthesis and aberrant mTOR activation through various mechanisms is frequently observed in a large portion of MM patients, contributing to cell survival, growth and drug resistance [12,13,14,15]. Moreover, accumulating research provides evidence that targeting the mTOR pathway can restrict protein synthesis in MM, resulting in cell death. Therefore, protein synthesis in general and the mTOR pathway specifically both represent interesting (new) targets in MM. This review will provide an update on what is known about the dysregulation of the mTOR pathway in MM and discuss promising new therapeutic strategies.

Overview of the mTOR pathwayStructure of the mTOR complexes

TOR is an evolutionarily conserved Ser/Thr-protein kinase that exists in two structurally and functionally distinct complexes, namely mTOR complex 1 (mTORC1), sensitive to the macrolide fungicide rapamycin, and the insensitive mTORC2 complex. They are both large complexes composed of multiple proteins. A regulatory-associated protein of mTOR (Raptor) and proline-rich AKT substrate 40 kDa (PRAS40) are specific to mTORC1, whereas mammalian stress-activated map kinase-interacting protein 1 (mSIN1), rapamycin-insensitive companion of mTOR (Rictor) and protein observed with rictor (Protor) 1 and 2 are exclusive components of mTORC2 (Fig. 1). However, they share mTOR, mammalian Lethal with Sec-13 protein 8 (mLST8), DEP-domain containing mTOR-interacting protein (Deptor) and the Telomere maintenance 2 (Tel2) and Tel2 interacting protein 1 (Tti1) complex.

Fig. 1

Schematic representation of the mTOR signaling pathway. mTORC1 and mTORC2 share mTOR, Deptor, mLST8, Tel2 and Tti1, while Raptor and PRAS40 are unique for mTORC1 and Rictor, mSIN1, and Protor are unique for mTORC2. Growth factors stimulate PI3K to convert PIP2 to PIP3. PIP3 will then recruit PDK1, leading to phosphorylation of AKT. In addition, RAS signaling can also be activated by growth factors, promoting the activation of RAF/MEK/ERK pathway. Activated AKT and/or ERK will then phosphorylate the TSC complex and/or PRAS40, leading to the relief of their mTORC1 inhibitory activity. For the TSC complex, phosphorylation by AKT will inhibit its GAP activity towards Rheb, allowing GTP-bound Rheb to bind to and activate mTORC1. Amino acids stimulate mTORC1 by promoting the formation of Rags-v-ATPase-Regulator complexes. In addition, Gln and Asn activate mTORC1 in a RAG-independent manner via the small GTPase Arf1. In contrast, energy stress will suppress mTORC1 activity by activating AMPK, resulting in the subsequent inhibition of Raptor and activation of the TSC complex. In addition, HIF-1 will prevent mTORC1 activation by inducing BNIP3 and/or REDD1, leading to Rheb inactivation. As for mTORC2, growth factors directly phosphorylate mSIN1 in a PIP3-dependent manner or through partially activated AKT, thereby promoting mTORC2 activation. Gs-coupled β2-adrenoceptor also promotes mTORC2 activation, by stimulating cAMP accumulation and PKA activation. In addition, AMPK directly activates mTORC2. In contrast, mTORC1 inhibits mTORC2 activation, by negatively regulating PI3K/AKT signaling through S6K1. mTORC1, mTOR complex 1; mTOR, Mammalian target of rapamycin; Raptor, Regulatory-associated protein of mTOR; Deptor, DEP-domain containing mTOR-interacting protein; PRAS40, Proline-rich AKT substrate 40 kDa; mLST8, Mammalian Lethal with Sec-13 protein 8; Rictor, Rapamycin-insensitive companion of mTOR; mSIN1, Mammalian stress-activated map kinase-interacting protein 1; Protor, Protein observed with rictor; Tel2, Telomere maintenance 2; Tti1, Tel2 interacting protein 1; PI3K, Phosphatidylinositol-3-kinase; PIP2, phosphatidylinositol-4,5-bisphosphate; PIP3, Phosphatidylinositol-3, 4, 5-triphosphate; PDK1, Phosphoinositide-dependent kinase 1; AKT, Protein kinase B; MEK, Mitogen-activated protein kinase; ERK, Extracellular-signal-regulated kinase; TSC, Tuberous sclerosis; GAP, GTPase-activating protein; Gln, Glutamine; Asn, Asparagine; Arf1, ADP-ribosylation factor 1; AMPK, Adenosine 5'-monophosphate-activated protein kinase; HIF-1, Hypoxia inducible factor 1; BNIP3, BCL2-interacting protein 3; REDD1, DNA damage inducible transcript 4; S6K1, Ribosomal S6 kinase; PKA, cAMP-dependent protein kinase

As a subunit of mTORC1, Raptor plays a crucial role in controlling the stability, lysosome surface localization, substrate recognition and function of mTORC1 [16,17,18,19,20]. By contrast, PRAS40 is recognized as an intrinsic inhibitory component of mTORC1, which binds to Raptor and competes with other substrates for mTORC1 binding, thereby inhibiting downstream signaling [21,22,23].

While mTORC1 has been well characterized in the last decade, knowledge on mTORC2 is only now rapidly developing. As a central member of the mTORC2 complex, mSIN1 contains an N-terminal domain (NTD), a RAS-binding domain (RBD), a conserved region in the middle (CRIM), and a pleckstrin homology (PH) domain in its C-terminal region. Both the RBD domain, through its interaction with active RAS, and the PH domain account for mTORC2 activation [24, 25], while the CRIM domain is in charge of mTORC2 substrate recruitment [26,27,28]. In addition, mSIN1 directly interacts with Rictor through its NTD, connecting Rictor with mLST8 to stabilize the mTORC2 complex [28, 29]. Rictor has comparable functions as Raptor, controlling mTORC2’s assembly, stability, and activity [30], whereby its C-terminal domain is responsible for mTORC2’s insensitivity to rapamycin [28]. Protor consists of two isoforms which also interact with Rictor through a conserved N-terminal region [31, 32], however, their role remains unclear.

When evaluating the shared components, mLST8 appears to be more important for the mTORC2 complex than the mTORC1 complex. Knockdown of mLST8 blocks activation of the mTORC2 substrates, while retaining the ability to phosphorylate mTORC1 substrates [33]. Studies indicate that this is mediated by interacting with the mTORC2 cofactors Rictor and mSIN1, thereby enhancing the assembly of the complex [34]. The stabilizing proteins Tel2 and Tti1 constitutively interact with mTOR in both mTORC1 and mTORC2, and the knockdown of either Tti1 or Tel2 results in the disassembly of both complexes [35]. Finally, Deptor is a highly conserved protein that binds to mTOR through its PDZ domain, thereby inhibiting the activity of both mTORC1 and mTORC2. However, Deptor and mTOR can also regulate each other, whereby mTOR kinase activity will phosphorylate Deptor, thereby promoting its release from mTOR and reversing its activity [36].

Regulation of the mTOR complexes

The activity of mTORC1 is regulated by several factors, including growth factors, amino acids, stress signals and cellular energy (Fig. 1). Several growth factors can activate mTORC1 by interacting with their cell-surface receptor tyrosine kinase(s), leading to the activation of the phosphatidylinositol-3-kinase (PI3K)/AKT and RAS/ERK (extracellular-signal-regulated kinase) pathways [37, 38]. By blocking either the tuberous sclerosis (TSC) complex or PRAS40, two mTORC1 negative regulators, AKT and ERK both positively control mTORC1 activity [39,40,41]. The TSC complex, which consists of three core subunits, TSC1, TSC2, and TBC1D7, keeps the small G-protein Rheb in an inactive state via its GTPase-activating protein (GAP) activity and by promoting Rheb ubiquitination [42, 43]. However, upon growth factor stimulation, AKT will phosphorylate both TSC2 and the deubiquitinase ubiquitin specific peptidase 4 (USP4), resulting in the release of Rheb from the inhibitory effect of the TSC complex [44]. PRAS40 is not only a component of mTORC1, but also a substrate of mTORC1, located downstream of mTORC1 but upstream of its effectors. Therefore, it can be controlled by both AKT or mTORC1 itself. While activated AKT dissociates PRAS40 from the mTORC1 complex by phosphorylating its threonine residue (Thr246), mTORC1 directly phosphorylates PRAS40 at serine residues (Ser183 and Ser221) to impair its inhibitory action [45,46,47].

It is generally believed that amino acid signaling stimulates mTORC1 activity by regulating its subcellular localization, and Rag guanosine triphosphatases (Rags or Rag GTPases) play a crucial role in this process [48, 49]. When amino acids are sufficiently present, active Rags form a complex with v-ATPase-Regulator and transmit amino acid signaling to the mTORC1 pathway by binding to Raptor. This process recruits mTORC1 to the lysosomal membranes, where Rheb is present, and stimulates mTORC1 activation [50, 51]. While most amino acids activate mTORC1 through Rags, glutamine (Glu) and asparagine (Asn) appear to activate mTORC1 in a Rag-independent manner that requires the small GTPase ADP-ribosylation factor 1 (Arf1) [52]. However, the glutamine sensor and other components involved in this Rag-independent pathway in mammals remain to be studied.

Energy stress controls mTORC1 activation primarily through an adenosine 5'-monophosphate-activated protein kinase (AMPK)-dependent mechanism. Under energy stress, such as glucose deprivation, the concentration of ATP drops dramatically while the cellular levels of AMP and ADP increase. AMP binds to the γ-subunit of AMPK contributing to its activation. AMPK then transmits the energy stress signal to mTORC1 mainly through two mechanisms [41, 53]. Firstly, AMPK activates the TSC complex, which in turn represses Rheb, thereby reducing mTORC1 activity [54, 55]. Secondly, AMPK will directly phosphorylate mTOR and Raptor, which also appears to be required for energy stress-induced inhibition of mTORC1 [56,57,58]. Additionally, AMPK-independent mechanisms have also been discovered to regulate mTORC1 activity upon stress. For example, mTORC1 can also be inactivated by hypoxia inducible factor 1 (HIF-1), the master regulator of the cellular response to hypoxia. HIF-1, either by inducing BCL2-interacting protein 3 (BNIP3) or by activating DNA damage inducible transcript 4 (DDIT4/REDD1), prevents activation of mTORC1 via direct interaction with Rheb [59,60,61,62].

In comparison to mTORC1, the signals activating mTORC2 and the mechanisms involved are less understood and more complicated. Similar to mTORC1, it is generally believed that growth factor-dependent mTORC2 activation requires PI3K/PIP3. In the unstimulated state, the mSIN1 PH domain is bound to the catalytic core within mTOR, thereby impairing mTORC2 activity. Following growth factor stimulation, PIP3 not only recruits Phosphoinositide-dependent kinase 1 (PDK1) and AKT from the cytosol, it will also bind to mSIN1 to expose the catalytic core within mTOR. AKT, which is partially activated through phosphorylation of Thr308 by PDK1, will then phosphorylate mSIN1 at Thr86, leading to a conformational change and subsequent promotion of mTORC2 activity. mTORC2 will then on its turn phosphorylate AKT at Ser473, resulting in full AKT activation [63, 64]. Additional stimuli that can trigger mTORC2 activation include adrenergic signaling via G-protein coupled receptors (GPCR), such as the β2‐adrenoceptor, which stimulates cAMP accumulation and activation of cAMP-dependent protein kinase (PKA), leading to phosphorylation of mTORC2 [65]. Also, AMPK appears to be sufficient to increase mTORC2 catalytic activity towards AKT in an mTORC1-independent manner [66]. Finally, mTORC2 activity is negatively regulated by mTORC1. Elevated mTORC1 activity upon insulin/ IGF-1signaling increases the activity of one of its direct effectors, S6K1 (see below), which in turn will phosphorylate insulin receptor substrate 1 (IRS1) on various negative regulatory sites, thereby inhibiting PI3K signaling and dampening mTORC2 [67].

Molecular mechanisms of mTOR-mediated translational control

mTOR functions as a central coordinator of cellular metabolic homeostasis in response to nutrient levels and growth signals. When ample nutrients and growth factors are present, the activation of the mTOR pathway promotes anabolic pathways, including protein and lipid synthesis, while also stimulating glycolysis and mitochondrial metabolism. Conversely, under conditions of hypoxia or energetic stress, mTOR signaling is inhibited, halting energy-consuming anabolic pathways and promoting catabolic pathways, such as autophagy [68]. In this review, we will discuss how mTORC1 and mTORC2 are involved in multiple aspects of protein synthesis, including activation of the substrates involved in mRNA translation initiation and promotion of ribosome biogenesis (Fig. 2).

Fig. 2

mTOR signaling and regulation of mRNA translation. mTOR signaling controls protein synthesis via regulation of mRNA translation initiation and ribosome biogenesis. mTORC1 phosphorylates 4E-BP1, resulting in the assembly of the eIF4F translation initiation complex. In addition, mTORC1 will phosphorylate S6K1, thereby promoting translation via phosphorylation of rpS6, eIF4B, PDCD4, eIF3, SKAR, and eEF2. In addition, mTORC1 also regulates ribosome biogenesis by activating UBF and TIF-1A, while inhibiting MAF1, thereby modulating Pol I and Pol III transcription. In addition, mTORC1 promotes translation of 5'-TOP transcripts by phosphorylating LARP1. Finally, mTORC2 also regulates ribosome biogenesis by relocating Rictor to the ER. mTOR, Mammalian target of rapamycin; 4E-BP1, Eukaryotic translation initiation factor 4E-binding protein 1; eIF4F, Eukaryotic translation initiation factor 4F; S6K1, Ribosomal S6 kinase 1; mTORC1, mTOR complex 1; rpS6, Ribosomal protein S6; eIF4B, Eukaryotic translation initiation factor 4B; PDCD4, Programmed cell death protein 4; eIF3, Eukaryotic translation initiation factor 3; SKAR, S6K1 Aly/REF-like substrate; eEF2, Eukaryotic elongation factor 2; Pol I/III, RNA polymerase I/III; LARP1, La-related protein 1; mTORC2, mTOR complex 1; Rictor, Rapamycin-insensitive companion of mTOR; 5’-TOP, 5’-terminal oligopyrimidine; ER, Endoplasmic reticulum

Activation of mRNA translation initiation

When sufficient nutrients are present, mTORC1 is strongly activated, promoting protein synthesis by phosphorylating eukaryotic initiation factor 4E-binding protein 1 (4E-BP1) and p70 S6 kinase 1 (p70-S6K, also known as S6K1) in a Raptor-dependent manner [69].

To initiate mRNA translation, the mRNA first needs to be unwound or activated by the eIF4F complex, comprising the cap-binding protein eukaryotic translation initiation factor 4E (eIF4E), the RNA helicase eIF4A, and the scaffold protein eIF4G, together with the assistance of eIF4B, eIF3 and poly(A)-binding protein (PABP). In its unphosphorylated state, 4E-BP1 represses translation by binding to and sequestering eIF4E, thereby preventing its interaction with eIF4G. mTORC1 phosphorylates 4E-BP1 at several sites, causing the dissociation of 4E-BP1 from eIF4E [70]. The release of eIF4E enables association with eIF4G and the assembly of the eIF4F translation initiation complex at the 5′end of the mRNA [71]. S6K1 is the second well-established downstream effector of mTOR that is directly phosphorylated by mTOR [72, 73]. S6K1 phosphorylates several factors participating in protein synthesis, including eIF4B, programmed cell death protein 4 (PDCD4), eIF3, eEF2, 40S ribosomal protein S6 (rpS6), and S6K1 Aly/REF-like target (SKAR). Phosphorylation of eIF4B leads to its binding with eIF4G and eIF4A, while phosphorylation of PDCD4 leads to its release from eIF4A, allowing eIF4A to interact with eIF4G. Importantly, eIF4B and PDCD4 phosphorylation by S6K1 is sufficient to maintain protein synthesis, even in the absence of 4E-BP1 [74]. Phosphorylated eIF3 will bind to the PABP regulatory protein PABP-interacting protein 1 (Paip1), thereby stabilizing the interaction between PABP and eIF4G, thus further stimulating translation [75, 76]. The protein kinase eukaryotic elongation factor 2 kinase (eEF2k) is a negative regulator of eEF2, which becomes inhibited after phosphorylation by S6K1, thereby releasing eEF2 and allowing proper elongation [77]. Phosphorylation of rpS6 has been shown to control cell size, however its function in protein synthesis remains elusive [78]. Finally, by interacting with SKAR, S6K1 is recruited to newly synthesized mRNAs in a splicing-dependent manner [79].

Ribosome biogenesis

To cope with increased protein synthesis, mTORC1 also promotes several steps in ribosome biogenesis, including ribosomal RNA transcription, synthesis of ribosome proteins and other components required for ribosome assembly. In mammals, the ribosomes contain 4 different rRNAs involved in ribosome assembly, which are transcribed by either RNA polymerase I (Pol I) or RNA polymerase III (Pol III) [80]. Several basal factors required for Pol I-mediated transcription are regulated by mTORC1. Firstly, mTORC1 activates Pol I-mediated transcription by increasing the expression and phosphorylation of UBF, thereby facilitating the recruitment of Pol I to rDNA [81]. Secondly, mTORC1 activates TIF-1A, a transcription factor that connects Pol I with UBF to initiate the transcription of pre-ribosomal RNA [82]. Thirdly, MAF1 is a key repressor of Pol III transcription, which becomes inhibited after phosphorylation by mTORC1 [83]. In addition, mTORC1 also controls the translation of a variety of mRNAs, particularly the 5’-terminal oligopyrimidine (5’-TOP) transcripts encoding ribosomal proteins, via direct phosphorylation of the La-related protein 1 (LARP1), a repressor of ribosomal protein mRNA translation [84]. Phosphorylation of LARP1 abolishes its blockage on the assembly of the eIF4F complex [85, 86]. Of note, enhanced ribosome biogenesis facilitates the transition of cells from an epithelial to a mesenchymal state, a process known as epithelial-mesenchymal transition (EMT). This EMT-associated ribosome biogenesis is accompanied by a pronounced increase in Rictor’s localization in the endoplasmic reticulum (ER), indicating also a regulatory role of mTORC2 in ribosome biogenesis [87].

Aberrant mTOR pathway signaling in MM cells

Over the years, dysregulation of mTOR has been associated with many diseases, such as diabetes, neurological disorders, and cancer (including MM) [88]. mTOR signaling is influenced in MM by numerous factors (Fig. 3), which can be subdivided in extrinsic, BM microenvironment-derived factors and intrinsic, cell-autonomous factors.

Fig. 3

Extrinsic and intrinsic factors regulating mTOR signaling in MM. Extrinsic factors: The myeloma growth factors IL-6, VEGF and IGF-1, which are abundantly present in the BM microenvironment, all induce mTORC1 activation via PI3K/AKT signaling. In addition, cell–cell contact with BMSC and osteoblasts via RANK-RANKL binding also activates PI3K/AKT/mTOR signaling in the MM cells. In addition, Pim2 overexpression, triggered by cytokines or cell–cell contact, also leads to mTORC1 activation via phosphorylating TSC2, while the hypoxic microenvironment mediates mTORC1 activity by regulating lactate, PYCR1 and MAT2A levels. Finally, β2AR is also involved in mTOR activation. Intrinsic factors: Deptor overexpression in MM cells blocks the inhibitory effect of S6K1 on AKT, thereby activating mTORC2. In addition, (Epi)genetic alterations, such as RAS mutationsPTEN depletionoverexpression of G9a/GLP and epigenetic silencing of RASSF4, all support enhanced mTORC1 signaling. Additionally, UCHL directly promotes the assembly of eIF4F. In contrast, Fbxo9 overexpression suppresses mTORC1 signaling by selectively targeting Tel2 and Tti1 in mTORC1 for degradation, which again releases mTORC2 from the negative feedback loop with mTORC1, leading to its activation. To maintain a high rate of protein synthesis, eIF4E is overexpressed in MM. Overexpressed eIF4E in turn promotes protein synthesis by upregulating MYC. Moreover, ER stress, induced by this massive protein synthesis, suppresses mTORC1 signaling via upregulating NUPR1. IL-6, Interleukin 6; IGF-1, Insulin-like growth factor-1; BM, Bone marrow; mTORC1, mTOR complex 1; mTORC2, mTOR complex 2; BMSC, Bone marrow stromal cells; RANK, Receptor activator of nuclear factor-kB; RANKL, RANK Ligand; mTOR, Mammalian target of rapamycin; PYCR1, Pyrroline-5-carboxylate reductase 1; MAT2A, Methionine Adenosyltransferase 2α; PI3K, Phosphatidylinositol-3-kinase; AKT, Protein kinase B; 4E-BP1, Eukaryotic translation initiation factor 4E-binding protein 1; S6K1, Ribosomal S6 kinase; Deptor, DEP-domain containing mTOR-interacting protein; PTEN, Phosphatase and tensin homolog deleted from chromosome 10; GLP, G9a-like protein; RASFF4, Ras-association domain family member 4; Tel2, Telomere maintenance 2; Tti1, Tel2 interacting protein 1; Fbxo9, F-box only protein 9; mTORC2, mTOR complex 2; ER, Endoplasmic reticulum; NUPR1, Nuclear protein 1; β2AR: β2 adrenergic receptor; UCHL1: Ubiquitin C-terminal hydrolase L1; TSC2: Tuberous sclerosis; eIF4F, Eukaryotic translation initiation factor 4F; eIF4E, Eukaryotic translation initiation factor 4E

Extrinsic, BM microenvironment-derived factors

IL-6 and IGF-1, as prominent MM growth factors s