2. Sphingolipids in Renal Cancer
Sphingolipids represent a vast class of lipids characterized by the presence of a sphingoid backbone in their structure. They are important constituents of cellular membranes, but are increasingly acknowledged for their role as signaling molecules.
Ceramide is the main hub of the sphingolipid pathways (
Figure 1). It is produced either through the de novo synthetic pathway, the salvage pathway, or the hydrolytic pathway. Ceramide may then be phosphorylated to form ceramide 1-phosphate (C1P), deacylated to sphingosine, or condensed with phosphatidylcholine to give sphingomyelin or glucose/galactose to give cerebrosides
[15][16]. Ceramide is known as a pro-apoptotic molecule, and many commonly used chemo-therapeutic agents induce cancer cell apoptosis by activating the acid sphingomyelinase and increasing ceramide formation
[17]. Additionally, in RCC cells, exposure to exogenous C6-ceramide, or increasing endogenous ceramide by a ceramidase inhibitor, had a cytotoxic effect
[18].
Figure 1. Sphingolipid biosynthesis and degradation routes. Changes reported for RCC are highlighted in pink boxes. For abbreviations, see text.
Phosphorylation of sphingosine by the two sphingosine kinases (SphK1 and SphK2) yields sphingosine 1-phosphate (S1P), which is a potent bioactive lipid involved in processes such as proliferation, migration, angiogenesis, lymphocyte trafficking, and endothelial permeability
[15][19]. Although a few intracellular targets of S1P are described, S1P is mostly known for its autocrine and paracrine functions through the five G protein-coupled receptors called S1P
1–5. Due to the anti-apoptotic and pro-angiogenic roles of S1P, the SphK/S1P/S1PR axis attracts special interest in cancer treatment
[20]. More specifically, this axis seems to be a master regulator of hypoxia by regulating HIF-1α and HIF-2α protein levels in human cancer cell lines including VHL-deficient ccRCC
[21].
In this regard, SphK1 inhibition decreases HIF-1α levels by stimulating its degradation in a pVHL-dependent manner. When pVHL is deficient as in the RCC10 cell line, HIF-1α levels are constitutively high and cannot be influenced by hypoxia or SphK1 inhibition
[22]. Moreover, SphK1 activity also controls HIF-2α expression and transcriptional activity, as SphK1 silencing promotes a VHL-independent HIF-2α loss which results in reduced cell proliferation in ccRCC
[23].
Interestingly, HIF-2α is also capable of regulating SphK1 levels by stimulating its gene transcription, protein expression, and enzyme activity. This is followed by increased intracellular S1P production, S1P release, and S1P receptor activation
[24]. According to The Cancer Genome Atlas (TCGA) RNA seq database
[25], SphK1 expression is 2.7-fold higher in solid tumor tissue from ccRCC patients, and S1P is increased in RCC tissue compared to healthy tissue
[26]. This is associated with poor survival and contributes to the resistance to the multi-kinase VEGFR inhibitor sunitinib
[25][27]. Mechanistically, it might involve the increased invasion mediated by S1P
2-dependent FAK phosphorylation
[25], increased viability, and proliferation through Akt/mTOR
[27], as well as an S1P
1+3-mediated increase in angiogenesis
[25]. Thus, the SphK1/S1P/S1PR axis is involved both in autocrine signaling to promote tumor growth, as well as in paracrine signaling to augment angiogenesis. Consequently, SphK1 inhibition was suggested as a possible strategy to control tumor hypoxia and its consequences
[28].
In order to prevent the stimulation of S1PRs and downstream signaling, antibodies against S1P
[29] or an extracellularly acting recombinant S1P lyase from
Symbiobacterium thermophilum (stSPL)
[30] were developed. Using the CAM model, Huwiler et al. (2011) demonstrated that this stSPL can indeed reduce tumor cell-induced angiogenesis. However, the short half-life in vivo, resulting in the rapid recovery of plasma S1P upon intravenous injection of stSPL in mice, hindered its further development.
Based on highly promising data of monoclonal murine (LT1002, Sphingomab) and humanized (LT1009, Sonepcizumab) S1P-specific antibodies in various preclinical models (cancer cell lines and in the retinal and choroidal neovascularization models in mice)
[31][32][33], clinical studies were initiated. Two formulations of sonepcizumab were designed, one for an intravitreous application (iSONEP) to treat exudative age-related macular degeneration and one for an intravenous infusion (ASONEP) to treat metastatic RCC. Despite the good safety and tolerability profile, sonepcizumab did not reach the primary endpoint of progression-free survival and the study with refractory RCC was terminated
[34]. Several limitations must be noted, such as the small patient number and a median of two prior failed therapies against VEGF/VEGFR, mTOR, or immunotherapy, which might anticipate a shorter progression-free survival
[34]. On the other hand, the encouraging overall survival (>20 months in a heavily pretreated population) and the favorable safety profile of sonepcizumab suggested that this agent could be explored in combination with the currently approved agents for metastatic RCC
[34]. So far, no such clinical trials were initiated.
Fingolimod (FTY720) which is an approved immunomodulator for the treatment of multiple sclerosis, acts as an unspecific agonist of all S1P receptors, except S1P
2, and as a functional antagonist of S1P
1 [35]. In several HIF-2α-resistant ccRCC cell lines, as well as in HIF-2α-resistant mouse ccRCC models, FTY720 showed anti-proliferative and anti-tumor effects
[21]. Whether the modulation of all S1PRs or only one of them is needed for these actions is not well understood. Fischl et al. (2019) recently reported that S1P
1 antagonism is sufficient to enhance the efficacy of the VEGFR inhibitor sunitinib in vitro and in vivo in the postnatal retinal angiogenesis model and in the RCC murine tumor model
[36]. This combination not only disrupted the tumor vascular beds, but also decreased the tumor volume and increased tumor cell death compared with monotherapies
[36].
On the other hand, siponimod, which is an S1P
1+5 modulator, was devoid of anti-proliferative effects in RCC colony formation assays, which was attributed to the inability of siponimod to target the S1P
3 receptor
[21]. Notably, RNA sequencing of five human ccRCC cell lines (769-P, A498, 786-O, SLR22, and RCC4) revealed that different cell lines expressed different levels of the five genes, with S1P
1 and S1P
3 being the most abundantly expressed subtypes
[21]. In RCC, the increased expression of insulin-like growth factor 2 mRNA-binding proteins enhances the stability of S1P
3 mRNA promoting cell proliferation and migration
[37], while patients with RCC characterized by a high expression of S1P
3 have significantly worse overall survival
[38]. These findings highlight the importance of a more selective approach when targeting the S1P receptors and more specifically of the potential of S1P
3 antagonists in RCC treatment.
The S1P
2 also deserves special attention, as this receptor subtype mediates an upregulation of connective tissue growth factor (CTGF) in the human cRCC cell line A498 following d16:1 S1P stimulation
[39]. In this study, it was reported that d16:1 S1P modulates conventional d18:1 S1P signaling by acting as a more potent agonist at the S1P
2 than the d18:1 S1P. It must be noted that all sphingoid bases, including d16, d18, and d20 chains, are produced in a rate-limiting step catalyzed by the same enzyme, the serine palmitoyltransferase (SPT)
[40].
Recently, it was reported that the decreased expression of one of the two major subunits of SPT, SPTLC1, predicts a poorer outcome in ccRCC patients and is significantly associated with disease progression
[41]. Moreover, SPTLC1 was decreased in RCC tissues compared to non-tumor tissues. The forced expression of SPTLC1 could significantly inhibit cell growth in vitro and in vivo in a nude mice xenograft RCC model via, at least in part, modulating Akt/FOXO1 signaling pathway
[42].
Interestingly, although the SPTLC1 mRNA levels decrease with the increase of the ccRCC stage
[41], which might suggest a reduced production of sphingoid bases, the content of dihydrosphingosine increased progressively with the increasing malignancy grade
[26]. Moreover, the level of dihydroceramide, which is an immediate precursor of ceramide, was elevated in G4 tumors, but not in lower malignancy grades
[26].
On the contrary, ceramide content, which is at a higher level in ccRCC than in non-cancerous kidney tissues, remained stable in tumors of higher malignancy grades despite the accumulation of dihydrosphingosine and dihydroceramide
[26]. This indicates either a block in the ceramide synthesis or a shift towards a particular ceramide subspecies. Notably, mammals have six ceramide synthases (CerS1-6), each exhibiting a preference for the chain length of the fatty acyl-CoA substrate and producing a distinct ceramide species
[43]. Indeed, data from RNA-seq databases show that RCC tumors exhibit increased CerS2 mRNA, which is inversely correlated with CerS6 mRNA in ABCB1+ clear cell carcinomas compared to normal tissue
[44]. Lipidomics analysis also showed a shift to predominantly longer chain ceramide and sphingomyelin species in chemoresistant ABCB1
high cells
[44].
The abovementioned study by Młynarczyk et al. (2022)
[26] reported on the expression levels of the major S1P-degrading or exporting factors, i.e., SPL, SPP1, SPP2, Spns2, and ABCC1, but no clear trend for a malignancy grade-depended expression was detected. Interestingly, the silencing of Spns2 blocked HIF-2α accumulation in ccRCC cell lines, thus mimicking the effect of the anti-S1P antibody
[23] and again highlighting the importance of S1P signaling in the regulation of HIF-2α as a driver of a more aggressive disease in RCC.
3. Glycosphingolipids in Renal Cancer
Galactosylceramide, or galactocerebroside, is produced from ceramide by the attachment of a galactose residue at the 1-hydroxyl moiety. α-Galactosylceramide (KRN-7000, α-GalCer) is a synthetic glycosphingolipid which acts as a synthetic iNKT (invariant Natural Killer T) cell ligand when presented by CD1d
[45]. This interaction activates the iNKT and increases the number of iNKT and the production of pro-inflammatory cytokines which later activate the NK, tumor-specific, CD4+, CD8+ T cells, and B cells
[46]. Numerous clinical trials have demonstrated tumor regression, a stable disease, partial response, or increased median survival time with α-GalCer therapy in various cancers; however, studies in RCC are missing
[46]. Although this immunotherapeutic vaccine approach was suggested to be of benefit in RCC
[45], the efficacy is unclear due to contradictory results and scarce studies. In this regard, in vitro α-GalCer-loaded dendritic cells induced the proliferation of iNKT cells derived from a pediatric papillary RCC
[47]. However, NK T cells isolated from peripheral blood mononuclear cells (PBMCs) of a fraction of patients with metastatic (m)RCC showed no functional activity towards autologous tumor cells in the presence of α-GalCer
[48].
Galactosylceramide is used by cerebroside/galactosylceramide–sulfotransferase (CST) to produce sulfatide. An elevated expression of sulfatide is commonly found in many human cancer cell lines and tissues and may possibly be used as a biomarker of some cancer cells
[49]. Sulfatide is a major L-selectin ligand in the kidney, and the binding between L-selectin and sulfatide plays an essential role in monocyte infiltration into the kidney interstitium
[49]. In various RCC cell lines, a marked increase of CST mRNA and activity was observed
[49]. Moreover, lactosyl- and galactosylceramide sulfate are markedly increased in RCC as compared to healthy tissue, accompanied by significantly elevated activities of their respective sulfotransferases
[50][51]. This is also reflected in the plasma and urine of RCC patients, where elevated concentrations of lactosylsulfatides were stage-dependent and more emphasized in late-stage RCC
[52]. Nevertheless, Porubsky et al. (2021) could not confirm an association between CST expression and malignant clinical behavior of RCC
[53]. Thus, the role of sulfoglycosphingolipids in RCC beyond the potential role as biomarkers for early RCC diagnosis
[54] at this moment lacks evidence.
The glucosylceramide synthase (GCS) is overexpressed in metastatic breast carcinoma
[55] and drug-resistant breast, ovary, cervical, and colon cancer cells
[56]. GCS upregulation is also part of the genetic signature for the progression and metastasis of RCC based on the results of gene-expression profiling of human RCC tumor samples
[57]. Since an overexpression of GCS confers drug resistance and the suppression of GCS expression overcomes the resistance by enhancing drug uptake and ceramide-induced apoptosis in breast cancer cells
[55][58], this suggests a mechanism that should also be considered in RCC.
Glucosylceramide serves as a substrate for the lactosylceramide synthase to build lactosylceramide. In a xenograft mouse model of RCC a significant correlation between the increase in the mass of lactosylceramide and the tumor volume was detected, and inhibition of GCS and lactosyl-ceramide synthase activities led to a decrease in tumor volume
[59].
Starting from lactosylceramide, globosides can be formed by the attachment of sugar residues, and gangliosides by the attachment of sugar residues and sialic acid.
It is now generally accepted that gangliosides produced by cancer cells play a role in immune escape. In the context of RCC, it was demonstrated that explanted RCC tumors produce soluble gangliosides that inhibit the nuclear factor κB activation of co-cultured T cells
[60], sensitize T cells to activation-induced cell death
[61], or directly induce T-cell apoptosis by caspase activation
[62].
For instance, RCC patients present with increased apoptotic T cells compared with T cells from healthy donors, and the majority of those apoptotic T cells were GM2(+) which they acquired from tumor shedding
[63]. GM2 originating from RCC was also shown to promote T cell dysfunction by down-regulating cytokine production
[64]. Not only do RCCs display increased levels of the gangliosides GD1a, GM1, and GM2 as compared with cells of the normal kidney
[62][65], but they also synthesize disialogangliosides which seemingly promote the metastatic capabilities through a mechanism involving the formation of microembolisms
[66].
Disialosyl globopentaosylceramide (DSGb5) is a dominant ganglioside isolated from RCC tissues
[67] which binds to sialic acid-binding Ig-like lectin-7 (Siglec-7) expressed on natural killer (NK) cells, thereby inhibiting NK-cell cytotoxicity
[68]. Higher DSGb5 expression exhibits greater migration potential in ACHN RCC cells and is correlated with metastasis in RCC patients
[68][69]. Other gangliosides, such as GalNAc disialosyl lactotetraosylceramide
[70] and monosialosyl galactosylgloboside (MSGG)
[71][72], bring a higher risk of metastasis; however, the exact mechanisms are still not thoroughly investigated.
Unlike gangliosides, the globosides globotriaosylceramide (Gb3) and globotetraosylceramide (Gb4) are markedly reduced in ccRCC tissue as compared to healthy renal tissue, and they decrease progressively with increasing malignancy grade
[26].
There seems to be a connection between the ganglioside and globoside content in RCC cells driven by the action of the plasma membrane sialidase NEU3
[73]. NEU3 silencing in a human primary RCC cell line led to an increase in ganglioside content (e.g., GD1a, GM2, and GM3), and a decrease in the globoside Gb3 content
[73]. Moreover, the production of ganglio-series gangliosides was enhanced to the detriment of globo-series gangliosides, particularly MSGG
[73]. The effects of this silencing on RCC cell malignant phenotype and behavior were significant and involved drug resistance, invasive potential, and adhesion
[73]. Nevertheless, other mechanisms could still play a role in these findings, as an increase of GM3 simultaneous with a decrease of MSGG in the human RCC cell line ACHN following brefeldin A treatment was marked by growth suppression and correlated to the pattern observed in RCC cases with a more favorable prognosis
[74].
Considering the proposed functions of gangliosides in other tumors, such as binding to endothelial cells through carbohydrate–carbohydrate interactions, modulation of adhesion receptors, or the promotion of tumor-associated angiogenesis
[66], this opens new avenues of research in the roles of gangliosides in RCC progression. Altogether, gangliosides expressed on RCC tumors may be important markers of tumor progression and target antigens for immunotherapy.
4. Free Fatty Acids in Renal Cancer
4.1. Exogenous Uptake of Fatty Acids
Over the last decades, extensive studies have approached the effect of free fatty acids (FA), particularly of ω3-polyunsaturated fatty acids (PUFAs) on cancer cells, and many epidemiological studies support the idea of a correlation between dietary FA intake and the development of cancer
[75]. Traditionally, saturated FAs have long been considered harmful, whereas plant monounsaturated FAs (MUFAs) such as oleic acid and ω-3 PUFAs were associated with a lower cancer mortality (
Figure 2). However, systematic reviews reveal only weak epidemiological evidence for a clear protection by MUFAs and ω3-PUFAs
[76][77]. Certain studies even concluded that certain MUFAs and PUFAs can promote cancer development
[78][79][80][81][82].
Figure 2. Fatty acid metabolism. Changes reported for RCC are highlighted in pink boxes. For abbreviations, see text.
Data approaching specifically the effect of FA intake on RCC is also scarce and contradictory. In an in vitro study in an RCC cell line, it was shown that PUFAs, including docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), were reducing the invasive profile of cells by upregulating the tissue inhibitor of metalloproteinase (TIMP)-1
[83]. It was hypothesized that ω-3 PUFAs modulated TIMP-1 synthesis by competing with the ω6-PUFA arachidonic acid for cyclo-oxygenase activity. This was supported by detecting reduced prostaglandin E
2 (PGE
2) production upon the addition of exogenous DHA which in turn elevated TIMP-1 protein levels. In a case control study conducted in Italy, both PUFAs and MUFAs seemed to be protective
[84], while a study evaluating the situation in a U.S. population cohort reported an elevated RCC risk with the increased dietary intake of animal fat, saturated fat, oleic acid, and cholesterol
[85]. Furthermore, a pooled analysis of 13 prospective studies showed statistically significant positive associations in pooled age-adjusted models for intakes of total fat, saturated fat, monounsaturated fat, polyunsaturated fat, and cholesterol and the incidence of RCC. However, after adjusting for BMI, fruit and vegetable intake, and alcohol intake, the statistically significant association was no longer seen
[86]. Clearly more studies are needed to prove dietary FA intake as a risk factor for RCC development.
4.2. Regulation of Fatty Acid Signaling in Cancer
The regulation of oncogenic signaling by FAs has also been considered as a novel therapeutic approach in RCC. Depending on the chain length, FAs can either freely enter the cell or use transport proteins
[87][88] (
Figure 2). Alternatively, cell surface FA receptors exist, denoted FFARs
[89][90] which are subdivided into two groups, the long-chain FFARs (FFAR1/GPR40 and FFAR4/GPR120) and the short-chain FFARs (FFAR2/GPR43 and FFAR3/GPR41). All these receptors belong to the superfamily of GPCRs. Their involvement in cancer cell growth and progression are only now beginning to be unmasked, but it seems that the long-chain FFARs have a different role than short-chain FFARs
[90]. Various in vitro studies in prostate, breast, ovarian, and colon cancer cells reveal that dual FFAR1/FFAR4 agonists can reduce the proliferation and migration of cancer cells, supporting the usefulness of these receptors as pharmacological targets
[91][92][93][94]. However, so far, no reports are available for FFAR involvement in RCC. Clearly, it will be important to optimize such FFAR agonists for selectivity and potency when considering them for further development.
4.3. Altered De Novo FA Synthesis
Metabolic reprogramming occurs because of mutations in cancer genes and alterations in cellular signaling. In addition to alterations in glucose and glutamine metabolism, increased de novo FA synthesis, uptake, and the suppression of FA oxidation, which eventually leads to lipid droplet (LD) formation, have been recently shown to be a hallmark of the disordered intermediary metabolism in cancer cells
[95][96].
The fatty-acid synthase (FASN) is the key metabolic multi-enzyme that is responsible for the terminal catalytic step in FA synthesis (
Figure 2). FASN is present at high levels in most human malignancies, especially in gynecological, prostate, and colon cancers
[97][98][99], and it is correlates with a worse prognosis
[100]. Therefore, FASN is speculated to be a new therapeutic target in RCC.
In a first study, Horiguchi et al. showed increased FASN protein staining in immunohistochemical sections of RCC patients
[101]. Positive FASN protein expression was associated with increased tumor aggressiveness and was an independent predictor of shortened cancer-specific survival, suggesting that FASN could be a predictive indicator of disease prognosis
[101]. These data were later confirmed by another study
[102], which assessed the differential mRNA expression of FASN in 533 ccRCC samples and 72 adjacent normal samples from a TCGA cohort. The data showed significantly increased FASN mRNA in ccRCC samples when compared to normal samples, and the elevated FASN mRNA correlated with a poor prognosis and malignant biological behaviors of ccRCC
[102]. Similar data were also obtained by Yuan et al.
[103] by using Western blot analysis and immunohistochemical staining of RCC tissue sections for FASN.
FASN up-regulation and its association with a poor prognosis holds true for other cancer types as well
[104][105][106], making this a universal cancer feature and thus supporting its usefulness as a therapeutic target of ccRCC. Wettersten et al.
[107] revealed that in RCC, metabolic reprogramming is grade-dependent. Interestingly, they reported that the levels of shorter chain FFAs (6:0, 8:0, 9:0, 10:0, and 12:0) were decreased in a tumor grade-dependent manner and this was probably due to an increase in their utilization. Nevertheless, also in this study, FASN was found to be increased on a protein level in cancer tissue when compared to the adjacent nontumor tissue
[108]. A functional analysis of FASN in human ccRCC cells showed that down-regulation or overexpression of FASN significantly regulates ccRCC cell proliferation and migration by regulating EMT. Moreover, FASN inhibition also increased the apoptotic rate, decreased lipid droplet formation, and suppressed the mRNA expression of hub genes in EMT
[102]. On top of this, the pharmacological inhibition of FASN reduced the growth and invasiveness of renal cancer cells in vitro and in vivo. One possible mechanism could be the disturbance of cell membrane functioning by down-regulated Her2 and EGFR and downstream STAT3 signaling
[101], which was shown to play an important role in pancreatic cancer metastasis
[109]. The ability of FASN inhibition to suppress cancer cell growth was also proven in a cell line of a pediatric malignant rhabdoid kidney tumor
[110]. Additionally, a proteomic analysis of tissue samples of a Wilms tumor confirmed that the expression of FASN was significantly increased in the tumor tissues as compared to adjacent tissues and this was associated with a poorer prognosis
[111][112]. All these data suggest that FASN plays a key role in ccRCC carcinogenesis and that the FASN expression level could be equally used as a predictor of poor prognosis in both pediatric and adult renal tumors.
Other than the
de novo synthesis, other enzymes in the FA pathway are involved in the altered lipid metabolism of RCC such as altered FA activation, FA uptake, and the suppression of FA oxidation
[96]. Once synthesized, FAs need to be activated by conversion to FA acyl-CoA esters by the action of acyl-CoA synthases (ACS) before they are further processed. Depending on the chain length of the Fas, ACSs are divided into different classes, comprising the very long chain (ACSVL), long chain (ACSL), medium chain (ACSM), and short chain (ACSS) synthases. Thus, ACSL converts FAs of C8-C22 chain lengths into an activated form. It represents a group of five isoforms, denoted ACSL1,3,4,5,6
[113]. Increased levels of all these enzymes have been suggested to be involved in molecular processes driving cancer growth and progression
[113][114]. While a clear view on the relevance of all the ACSL isoforms in RCC is not yet available, it seems that at least ACSL3 could serve as a potential prognostic biomarker for immune infiltration in ccRCC
[115]. Furthermore, ccRCC cells in vitro depend on ACSL3 for lipid droplet formation. Selective pharmacological inhibition or the genetic suppression of ACSL3 is cytotoxic for RCC cells and also reduces the tumor size in an orthotopic mouse cancer model
[116]. These data propose that ACSL3 could indeed be a possible pharmacological target for RCC therapy.
5. Conclusions and Perspectives
The reprograming of lipid metabolism is a typical characteristic feature of many tumors. Many of the so-far-described bioactive lipids can interfere with cancer-relevant molecular processes including cell proliferation and migration, apoptosis or survival, angiogenesis, and metastasis formation. Therefore, their involvement in RCC is very obvious and this has been approached in multiple in vitro and in vivo studies over the last decades. Another hallmark of RCC is the up-regulation of the HIF/VEGF signaling axis. It turns out that the targeting of the HIF/VEGF axis is very efficient in RCC treatment. However, resistance development is a main problem, which stresses the need for better treatment options. Not surprisingly, many of the bioactive lipids are regulating HIF/VEGF signaling or the opposite, are themselves regulated by HIF/VEGF, which highlights the attractivity of the lipids as a new targeting strategy for RCC, and it will be exciting to see whether novel therapeutics can arise from these lipid cascades.