The mechanistic (formerly mammalian) target of rapamycin (mTOR) influences various cellular processes (proliferation, motility, migration, metabolism, protein synthesis, transcription, etc.,) by integrating signals from the tissue environment. mTOR promotes cell growth and survival, depending on the actual state of the cell. In addition, a decrease in mTOR activity in the absence of growth factors, nutrients, or other energy sources (e.g., oxygen) inhibits cell growth and triggers survival-promoting cellular processes such as autophagy. As a central signaling pathway hub, mTOR kinase plays an essential role in metabolic regulation to maintain the balance between anabolic and catabolic processes, including metabolic adaptation in stress responses and tumor cell survival (Figure 1A). Dysfunction (hyper- or underactivation) of the mTOR kinase may contribute to regulatory failures and subsequently to the development and progression of diseases (e.g., metabolic, neurodegenerative, and cardiovascular diseases, accelerated aging, tumors) [1, 2].

mTOR is a serine-threonine kinase that forms two complexes (mTOR complex 1—mTORC1; mTOR complex 2—mTORC2) organized from different, functionally distinct proteins (Figure 1B). The identical subunits in the two complexes are mTOR kinase, mLST8 (mammalian lethal with sec-13 protein 8), and Deptor (DEP domain containing mTOR-interacting protein). PRAS40 (proline-rich Akt substrate of 40 kDa) and the scaffolding protein Raptor (regulatory-associated protein of mTOR) are involved in the assembly of mTORC1. mSin1 (mammalian stress-activated map kinase-interacting protein 1), Protor 1/2 (protein observed with Rictor 1 and 2), and the scaffolding protein Rictor (rapamycin-insensitive companion of mTOR) instead of Raptor are in mTORC2 [3].

mTORC1 plays a role in several cellular processes (e.g., protein, lipid and nucleotide synthesis, ribosome biogenesis), whereas mTORC2 has a significant role in the modulation of cell survival, differentiation, growth, and migration and the maintenance of the actin cytoskeleton, mainly through the phosphorylation of Akt (Ser473), PKCα, and SGK1 [4].

mTOR complexes and their dysfunction—mainly hyperactivity—have been associated with the development and progression of malignancies. Hyperactivity of mTOR complexes can occur in several ways; including gene mutations affecting mTOR kinase, mutations in proteins that directly regulate the mTOR complex activity, or other changes in the signaling network (e.g., mutations in oncogenic or tumor suppressor genes).

Nearly 5% of solid tumors may carry an activating mutation of mTOR kinase, which may be more common (>5%) in melanoma, endometrial, gastrointestinal, kidney, breast, and lung cancers [5]. On the basis of publicly available databases, the frequencies of genetic alterations associated with/related to mTOR kinase in different malignancies are summarized in Figure 2.

The enzymatic activity and oncogenic role of mTOR kinase are generally increased by somatic mutations affecting six different kinase regions. The most common activating mutations are E1799K, T1977R, V2006F, S2215Y, and R2505P, which are responsible for amino acid substitutions (Figure 3). Activating mutations do not affect the assembly of the mTOR complex but can reduce the binding of Deptor—an inhibitor of mTOR kinase—and increase the phosphorylation of target molecules (e.g., P70S6K, 4EBP1) [6, 7]. Some mutations alter the structure of the complex, leading to resistance to allosteric mTOR inhibitors (e.g., rapamycin) by preventing drug binding. For instance, F2108L point mutation occurs in the FKBP-binding (FRB) domain of mTOR, where the FKB12-rapamycin binding site is located [8].

Most commonly, other mutations can cause pathway hyperactivity or loss of negative regulators in the signaling network (e.g., PI3KCA, PTEN, TSC1/2, STK11, and AKT1). PI3KCA mutations occur in more than 20% of breast and gynecological cancers and are also common in colorectal tumors. PTEN mutations are found in 10% of central nervous system and endometrial tumors, TSC1 mutations in 5%–6% of urinary tract and endometrial tumors, and TSC2 mutations in 4%–7% of gynecological, liver, and lung tumors. STK11 mutations occur in 10%–15% of the cervical, small bowel, and lung tumors, while AKT1 mutations are the least frequent of the signaling mutations (3%–5%). In many cases, mutations in receptor tyrosine kinases and growth factor receptors (EGFR and HER2) underline mTOR hyperactivity [9].

In addition to direct mutations in the mTOR kinase and the above-mentioned mutations affecting mTOR signaling, other gene mutations can occur in different subunits of the mTOR complex. One of the most common alterations is RICTOR amplification, which—in addition to increased expression of the Rictor protein—can increase the activity of the mTORC2 complex. In this context, it can affect/activate the target protein Akt kinase, which plays an essential role in the growth and survival of tumor cells (see below).

The discovery of mTOR complexes and the potential therapeutic relevance of the mTOR kinase inhibitor rapamycin focused attention on the need to characterize mTOR activity. First-generation allosteric mTOR inhibitors are effective in the case of mTORC1 hyperactivity, although these may be ineffective in specific known FRB domain mutations or the presence of high mTORC2 activity. Therefore, the development of a new generation of mTOR kinase inhibitors that inhibit both mTORC1 and mTORC2 and additional inhibitors such as dual inhibitors has been initiated. Based on many published data, it is necessary to characterize mTOR activity in situ before starting mTOR-targeted therapies, as the efficacy of drug treatment may also depend on the ratio of the complexes.

In our studies, we have developed marker panels that can be used in a wide range of tumor types to quantify and morphologically characterize not only mTORC1 but also mTORC2 complex activity in tumor cells or tissues. Characterization of mTOR activity in different cell lines and patient material can be performed by staining/measuring the active, phosphorylated form of mTOR kinase (p-mTOR) and its phosphorylated targets (e.g., p-4EBP1, p-p70S6K, p-S6) using immunohistochemistry (IHC)/immunocytochemistry and/or Western blot/Wes^TM Simple. One of the target proteins of the mTORC2 complex is the Akt kinase, where the amino acid serine 473 can only be phosphorylated by active mTORC2. On the basis of publicly available results, quantitative and in situ analysis of p-Akt (Ser473) is the most appropriate way to monitor the mTORC2 activity.

A quantitative comparison of the two mTOR complexes can also be determined from the ratio of specific scaffold proteins (Raptor for mTORC1 and Rictor for mTORC2). The optimal in situ marker panel is either a combination of p-mTOR, Raptor, and Rictor or a combination of mTORC1 target proteins (e.g., p-S6, p-70S6K, p-4EBP1) and a specific mTORC2 target protein [p-Akt (Ser473)] (Figure 4).

The first clinical anti-tumor observation of the mTOR inhibitor, the immunosuppressive drug rapamycin, was detected in the treatment of post-transplant kidney cancer [10]. Following the discovery of mTOR complexes, the mTOR hyperactivity and/or sensitivity of tumor cells against mTOR inhibitors, quantitative changes in the mTOR signaling pathway elements, target proteins, and their active forms have been described and partially characterized in various tumor types by the end of the last century and early 2000 s (detailed in Figure 5).

Our studies have contributed to the characterization of mTORC1 activity in hematological malignancies and even to the investigation of both mTOR complexes in leukemias and lymphomas. Summarizing these studies, acute lymphoblastic leukemia (ALL) cells and several lymphoma model cell lines (Hodgkin’s lymphoma, diffuse large B-cell lymphoma (DLBCL), anaplastic large cell lymphoma, mantle cell lymphoma, lymphoplasmacytic lymphoma, Burkitt’s lymphoma) show high mTORC1 activity. In addition to the generally high mTOR activity, the poorest survival outcomes were observed with the co-occurrence of high mTOR activity and increased Rictor expression. These findings highlight the unfavorable prognostic role of increased mTORC2 activity in the majority of lymphoid malignancies studied [38, 40, 42].

In addition to in vitro lymphoma models, mTOR hyperactivity has been characterized in the most common solid tumors and several human malignant tissues. In gastrointestinal tumors (colon tumors, gastrointestinal stromal tumors—GIST), we first characterized the activity of mTOR complexes, based on the amount of mTORC1 and mTORC2 scaffold proteins, and classified Hungarian colon tumor cases into three main groups (low Raptor and high Rictor expression—high mTORC2 activity; high Raptor and low Rictor expression—high mTORC1 activity; similar levels of Raptor and Rictor expression—balanced mTORC1 and mTORC2 activity). We compared our results with the clinical characteristics of the patients and found the best treatment and survival outcomes in cases with low mTORC1 activity, whereas the worst survival outcomes were observed in cases with high mTOR activity and high Rictor expression. Our results also showed that Rictor expression was associated with poor prognosis independently of mTOR activity [37, 44]. In our human breast cancer studies over the past few years, we have detected mTORC1 hyperactivity in 50% of our cases, supporting the use of previously introduced mTOR inhibitor therapy. In addition, we observed high Rictor expression—the increased presence of the mTORC2 complex—in more than 40% of the cases studied, which may explain the ineffectiveness of conventional rapalogue (rapamycin and its derivatives) treatments [48]. We also characterized mTOR activity in renal tumors. In our comparative studies, we could detect higher expression of p-mTOR and p-S6 markers in clear cell renal cell carcinoma compared to papillary renal cell carcinoma and normal tubular epithelial cells. Moreover, papillary renal cell carcinoma was characterized by high Rictor expression and potential mTORC2 activity, also highlighting the role of mTORC2 activity in this tumor type [51]. There has been an emerging need to characterize mTOR activity in rare solid tumors. Our studies also aimed to investigate the mTOR activity in rare tumor types, such as central nervous system tumors, childhood rhabdomyosarcoma, osteosarcoma, medulloblastoma, and fibrosarcoma [43, 49, 50, 52, 53].

Rictor is a characteristic actin coordinating scaffold protein of the mTORC2 complex, and its primary function is to ensure the formation and structure of the mTORC2 complex. Rictor contributes to the formation of a spatially structured mTORC2 complex in which the FKBP-rapamycin binding region of the mTOR kinase is located inside the molecule, blocking access to the best-known and most widely used inhibitor of rapamycin and its first-generation derivatives (rapalogs) [1]. Consequently, these inhibitors do not directly affect the mTORC2 target proteins (e.g., Akt). Increased Rictor expression is generally associated with transcriptional and translational changes. One of the best-known oncogenic mutations affecting the mTOR kinase pathway is the amplification of the RICTOR gene. RICTOR amplification can lead to increased expression of the Rictor protein and, in most cases, an increase in mTORC2 complex activity and a shift in the ratio of mTORC1 to mTORC2 complexes. RICTOR amplification has been described to play an essential role in the progression of certain tumors and metastasis via the regulation of signaling pathways (e.g., MAPK/ERK, Wnt/β-catenin pathways) [54].

Among the genetic defects of RICTOR, the most frequently observed change is its amplification, but other alterations have also been described [55]. The genetic alteration frequency of RICTOR in different malignancies is summarized in Figure 6 (based on public databases).

Additionally, higher Rictor expression has been described in several tumor types, including lung [56], esophageal [57], breast [58], pancreatic [59], liver [60], and melanoma [61], where high Rictor expression was associated with poorer prognosis and shorter survival [55]. Table 1 summarizes mTORC2 hyperactivity and/or RICTOR amplification in different tumor types.

RICTOR amplification may indicate high mTORC2 activity and may be a predictive marker of effective inhibition of the PI3K/mTOR/Akt axis. Fluorescence in situ hybridization (FISH) is considered the gold standard validation diagnostic method for identifying RICTOR amplification (Figure 7). In addition, potential changes in copy number can be detected by sequencing (e.g., next-generation sequencing) or Droplet Digital PCR (ddPCR). The Rictor protein expression can be detected by immunohistochemistry or immunocytochemistry. As mentioned above, the mTORC2 complex is exclusively responsible for the phosphorylation of serine 373 in the Akt protein. Increased p-Akt (Ser473) protein expression is a clear sign and an excellent marker of increased mTORC2 complex activity [46, 50]. It is important to note that tissue sample preparation and fixation require special attention when testing for p-Akt (Ser473) expression. Failures in the process and the timing of tissue sample fixation may reduce the detectability of phosphorylated proteins.

In our latest study, the diagnostic next-generation sequencing presumed RICTOR amplification was further analyzed in different human tumor tissues. RICTOR amplification was tested by ddPCR and validated using the “gold standard” FISH. In addition, Rictor and p-Akt (Ser473) protein expression was also examined by immunohistochemistry. The 10 novel and 4 previously described various tumor types with FISH-validated RICTOR amplification demonstrate the significance of RICTOR amplification in a wide range of malignancies (detailed in Figure 8). The newly described entities with RICTOR amplification may initiate further studies with larger cohorts to analyze the prevalence of RICTOR amplification also in rare diseases [97].

Dysregulation of the PI3K/Akt/mTOR signaling pathway plays an important role in the development and progression of the majority of lung cancers [98]. mTOR activity can be affected by failures in other signaling pathways related to the mTOR pathway and the known alterations in mTOR pathway activity (see above). In this context, the combined inhibition of mTOR and other kinases may provide additional therapeutic options for treating lung tumors [56].

Among non-small cell lung cancers (NSCLC), 90% of adenocarcinomas, 40% of squamous cell carcinomas, and 60% of large cell carcinomas are characterized by increased activity of PI3K/Akt/mTOR axis [99]. Elevated levels of phosphorylated mTORC1 targets (e.g., p-4EBP1, p-S6) have been observed in adenocarcinoma and squamous cell carcinoma. In these studies, tumor invasiveness, metastasis, and poorer prognosis are associated with increased activity of mTOR [100]. Activating mutations in PIK3CA have been detected in 4%–7% of NSCLCs, amplification in more than 30% of squamous cell carcinomas, and about 1% of adenocarcinomas. RICTOR amplification has been described in 10% of NSCLCs, and mutations in the tumor suppressor genes PTEN and STK11, which are negative regulators of mTOR signaling, have also been observed in squamous cell and adenocarcinomas [101–103]. Moreover, mTOR activation correlates with mutations in both KRAS and epidermal growth factor receptor (EGFR) and may serve as a resistance mechanism to treatment with EGFR inhibitors [104].

In our studies, we examined the p-mTOR, p-S6, and Rictor proteins in situ in primary lung adenocarcinomas and brain metastases. Increased p-mTOR, p-S6, and Rictor protein expressions were observed in more than 30% of primary adenocarcinomas and around 70% of brain metastases. Additionally, all studied markers (p-mTOR, p-S6, and Rictor) showed a stronger positivity in the majority of the brain metastases studied. In some cases, high p-mTOR levels were associated with low p-S6 and high Rictor expression, which characterized about 20% of primary lung adenocarcinomas and more than 50% of brain metastases. Further associations were found between higher stage and Rictor expression in primary adenocarcinomas and high Rictor and low p-S6 expression in solitary brain metastases. Comparing primary tumor and brain metastases samples from the same patient, we observed increased mTORC1 activity in metastases; in 60% of the cases, p-mTOR and p-S6 expression was increased in brain metastases, whereas in 40% of cases, Rictor expression was increased in brain metastases. These data highlight the presence and prognostic role of mTORC2 activity in this tumor type [45].

Among lung tumors of neuroendocrine origin, small cell lung carcinoma (SCLC) is the most common, accounting for 15%–20% of all lung tumors [105]. In addition to mutations in the TP53 and RB1 genes and MYC amplification, genetic alterations affecting elements of the PI3K/Akt/mTOR pathway (e.g., PIK3CA, PTEN, AKT2, AKT3, MTOR, and RICTOR) are common in SCLC. RICTOR amplification is the most common targetable genetic alteration in SCLC (6%–14%) [106]. Activation of the mTOR pathway in SCLC has been detected by immunohistochemistry for p-mTOR and p-S6 proteins [33], although less data are available on the activity, background, and significance of mTORC2.

Our RICTOR amplification studies were initiated in collaboration with the Mayo Clinic, and within the framework of this collaboration, we analyzed 100 samples from 92 SCLC patients. The presence of RICTOR amplification was confirmed by FISH. We detected RICTOR amplification in 15% of cases, 3% of our cases were equivocal, and 82% were negative. In these samples, we characterized Rictor and p-Akt levels by immunohistochemistry. Of the cases, 14% showed high, 23% moderate, 25% low Rictor expression, and 38% were negative. The expression of p-Akt was high in 16%, moderate in 26%, and low in 35% of the cases, whereas in 23% of the tested samples, we could not detect the presence of the protein. Statistical analyses showed that Rictor immunohistochemistry had a sensitivity of 93% and a specificity of 73%; p-Akt immunohistochemistry had a sensitivity of 80% and a specificity of 65% compared to RICTOR FISH results. The presence or absence of RICTOR amplification was not associated with the overall survival data of the patients. However, higher in situ expression of Rictor and p-Akt proteins was significantly associated with shorter overall survival [46].

A rare lung disease, lymphangioleiomyomatosis (LAM), is caused by a loss-of-function mutation in the TSC1/2 gene, resulting in lung tissue damage through proliferation, growth, and invasion of LAM cells [107, 108]. The therapeutic relevance of mTORC1 activity and the expression of downstream markers of the mTOR signaling pathway (e.g., p-p70S6K, p-S6, p-4EBP1) have been demonstrated by immunohistochemical studies [109, 110].

With the help of the Mayo Clinic, 11 cases of LAM were involved in our studies. Most of these (91%) showed high p-S6 protein expression, demonstrating increased mTORC1 activity and supporting mTOR inhibitor treatments. In parallel, high levels of Rictor expression were detected in addition to p-S6 in more than half of the cases (55%). This observation also highlights the role of the potential activity of the mTORC2 complex in this disease, which may explain the ineffectiveness of mTORC1 inhibitors in certain cases. This study was the first to examine both mTORC1 and mTORC2 activity in human LAM samples, and in addition to previously reported mTORC1 hyperactivity, we have also demonstrated increased mTORC2 complex presence and activity. Further studies demonstrating mTORC2 hyperactivity may suggest using other dual mTORC1 and mTORC2 inhibitors or dual mTOR inhibitors to slow the progression of this rare disease [47].

Our studies support the importance of characterizing mTOR activity in a wide range of tumors, which may indicate the clinical need for the appropriate use of mTOR inhibitors. In these cases, we recommend the immunohistochemical analysis of at least 3-4 in situ tissue markers [e.g., p-mTOR, p-S6, Rictor, and p-Akt (Ser473)] to provide a good representation of mTORC1 and mTORC2 activity in tumor tissue. It is important to note that first-generation mTORC1 inhibitors may be ineffective in the presence of high Rictor and p-Akt (Ser473) expression, as tumor cells with high mTORC2 activity may survive despite the administration of specific mTORC1 inhibitors [111].

Over the past decades, the efficacy of several PI3K/Akt/mTOR signaling inhibitors has been investigated. Despite numerous clinical trials with inhibitors of this signaling pathway, therapeutic responses are often lower than expected; moreover, only a few new drugs have been introduced to treat various tumors. One possible reason for this could be an inadequate patient selection method and, in this context, the lack of reliable predictive markers. Sensitivity to mTOR inhibitors may be predicted by specific genetic abnormalities (e.g., PIK3CA mutation/amplification, PTEN loss-of-function mutation, AKT mutation, RICTOR amplification) or overexpression of the signaling pathway elements and their active targets (e.g., Rictor, p-S6, p-Akt). These markers can be well assessed at the tissue level, and their presence or absence should be evaluated/considered when designing targeted therapies before specific inhibitors are used [112, 113].

The best-known mTOR inhibitor, rapamycin (or sirolimus), also the eponymous inhibitor of mTOR kinase, was discovered by George Nogrady, a Hungarian-born bacteriologist, in the Chilean territory of Easter Island (Rapa Nui) in 1964 [114]. The compound was isolated as a microbial antibiotic and introduced as an immunosuppressant in 1975 [115]. Following the discovery of the specific target protein of rapamycin, mTOR kinase, allosteric mTOR inhibitors, rapamycin and its derivatives (rapalogs—everolimus, temsirolimus, etc.,) have been used with varying results in several types of tumors (e.g., breast, kidney, endocrine, central nervous system). Rapalogs inhibit the activity of mTORC1 by binding to the FRB domain of mTOR kinase through the FKBP12 protein. Due to the structural differences between the two complexes (see above), no direct effect of rapamycin is observed for the mTORC2 complex [111]. Further studies are needed to determine the possible impact of rapamycin on the mTORC2 complex.

In addition to rapalogs, double mTORC1 and mTORC2 inhibitors (e.g., CC-115, sapanisertib, vistusertib), as well as dual mTOR kinase and another signaling kinase (e.g., PI3K) inhibitors (e.g., gedatolisib, paxalisib, samatolisib) are currently under development. Other specific Akt inhibitors are also being tested (e.g., afuresertib, capivasertib, ipatasertib, MK-2206, TAS-117, triciribine, uprosertib) and are currently in phase II and III trials. Third-generation mTOR inhibitors (e.g., RapaLink-1) are also being developed for the treatment of various advanced cancers (e.g., renal, breast, mantle cell, and other high-grade lymphomas) [116]. Table 2 summarizes PI3K/Akt/mTOR inhibitors currently in use and development.

Developing potent and highly selective small molecules specific to the mTORC2 complex while preserving mTORC1 activity remains challenging. Among the specific inhibitors of mTORC2, only a few inhibitors based primarily on siRNA technology (e.g., Rictor si-NP, JR-AB2-011) are under preclinical testing [64, 117].

Despite the association between PI3K/Akt/mTOR hyperactivation and adverse prognosis, the efficacy of mTOR inhibitors used as monotherapy is very low, as is the case for most targeted therapeutic agents. However, combining mTOR inhibitors with other targeted therapies or conventional chemotherapy/radiotherapy may help sensitize different agents and overcome resistance mechanisms to treatment. This type of combination therapy is also being developed for a wide range of malignancies (Table 3) [118]. However, the administration of these mTOR inhibitor combination therapies can also be limited by adverse effects and/or severe individual side effects that require very careful management by oncologists [119].

Advances in personalized therapies and the increasing availability of molecular genetic data on malignancies, including our findings of mTORC1 and mTORC2 activity and RICTOR amplification in various tumors, highlight the importance of validated targets in common and rare malignancies. The efficacy of targeted therapies often fails to deliver the expected results due to inadequate patient selection. In a biomarker-driven umbrella study, patients were selected based on RICTOR amplification, demonstrating a promising strategy for personalized therapies by selecting patients most likely to respond to targeted treatments. In this clinical trial, two of the four SCLC patients treated with a vistusertib (dual mTORC1 and mTORC2 inhibitor) showed an increase in survival of almost 1 year [120].

In our work, we have developed a marker panel to determine tissue characteristics of mTOR activity in biopsy specimens, patient materials, and cell lines. Our results and other molecular findings will hopefully support and optimize future therapeutic decisions.

mTOR hyperactivity can be a targeted genetic alteration in different tumor types. In our studies, we present our findings on the activity of both complexes and the characterization and prognostic role of mTORC2 complex hyperactivity. In the future, we intend to investigate further the significance of mTOR hyperactivity and RICTOR amplification in several malignancies. Expanding next-generation sequencing into routine diagnostics will facilitate more accurate oncogenic and tumor suppressor gene alterations mapping. Molecular pathology results and molecular genetics data will become widely available, which may help to identify mutations affecting mTOR activity. This will also help to determine the importance of specific mutations in rare tumor types.

In summary, several third-generation PI3K/Akt/mTOR pathway inhibitors are already developing, and some of these are in clinical trials. Only a few inhibitors have been approved for treating various cancers compared to the enormous amount of research and money spent on their development. To increase this number and to achieve clinical translation of more mTOR inhibitors or other targeted inhibitors into personalized therapies, it is essential to identify predictive markers that can help therapeutic decision-making. In conclusion, for biomarker-driven patient selection, it is crucial to develop agents that are as effective as possible while maintaining a good safety profile and use rational drug combinations that improve quality of life and may be effective in overcoming primary or acquired resistance.

The figures in this review (Figures 1, 4, and 6) were edited using BioRender (https://biorender.com) under license from the Department of Pathology and Experimental Cancer Research, Semmelweis University.