Although all cells depend upon nutrients they acquire from the extracellular space, surprisingly little is known about how nutrient uptake is regulated in mammalian cells. Most nutrients are brought into cells by means of specific transporter proteins. In yeast, the expression and trafficking of a wide variety of nutrient transporters is controlled by the TOR (target of rapamycin) kinase. Consistent with this, recent studies in mammalian cells have shown that mTOR (mammalian TOR) and the related protein, PI3K (phosphoinositide 3-kinase), play central roles in coupling nutrient transporter expression to the availability of extrinsic trophic and survival signals. In the case of lymphocytes, it has been particularly well established that these extrinsic signals stimulate cell growth and proliferation in part by regulating nutrient transporter expression. The ability of growth factors to control nutrient access may also play an important role in tumour suppression: the non-homoeostatic growth of tumour cells requires that nutrient transporter expression is uncoupled from trophic factor availability. Also supporting a link between nutrient transporter expression levels and oncogenesis, several recent studies demonstrate that nutrient transporter expression drives, rather than simply parallels, cellular metabolism. This review summarizes the evidence that regulated nutrient transporter expression plays a central role in cellular growth control and highlights the implications of these findings for human disease.
- mammalian target of rapamycin (mTOR)
- nutrient transporter
- phosphoinositide 3-kinase (PI3K)
- transferrin receptor
In 1972, Robert W. Holley, Nobel Laureate for his work demonstrating that tRNA interprets the genetic code, published a short “Unifying hypothesis concerning the nature of malignant growth” . In this article, he suggested that mammalian cell growth is fundamentally regulated at the level of intracellular nutrient availability. Holley hypothesized that, although the bloodstream supplies cells with constant levels of nutrients, hormones and/or growth factors regulate access to these nutrients by modulating the expression of nutrient transport systems in the cell membrane. Malignant growth would result from increased intracellular nutrient concentrations due to the uncoupling of nutrient transporter expression from trophic factor availability. This model contrasts with the widespread belief that nutrient transporter expression is part of a ‘housekeeping’ programme that simply meets the metabolic demands of the cell. In Holley's model, nutrient transporters function as gatekeepers, not housekeepers, by controlling cellular access to the fuel required for proliferation and growth.
The framework for this control mechanism may be borrowed from free-living unicellular organisms such as yeast, where growth and survival is generally nutrient-limited. Yeast use signal transduction pathways to match their nutrient transporter complement to the extracellular conditions and thus maximize their ability to proliferate. These nutrient-responsive signal transduction programmes are cell-autonomous: each yeast cell expresses an array of transporter proteins designed to allow maximal growth in the available nutrient milieu without regard for the needs of its neighbours. In contrast, mammalian cells exhibit co-operative homoeostatic growth, despite the fact that a constant supply of nutrients is delivered via the bloodstream. One of the restraints on mammalian cell growth comes from the fact that the nutrient-responsive signal transduction pathways regulating transporter expression in yeast are harnessed to extrinsic growth factor receptors in mammalian cells. In the absence of growth factors, nutrient transporter proteins are degraded, and mammalian cells starve to death even in rich media. In fact, growth withdrawal phenocopies nutrient deprivation: anabolic metabolism declines, the mitochondrial membrane potential falls, and cells atrophy, arrest in the cell cycle, rely on autophagy for energy and eventually initiate programmed cell death (apoptosis). By placing growth factors in charge of nutrient transporter expression, mammalian cells have lost the capacity for cell-autonomous growth, but have gained an efficient mechanism for tissue homoeostasis.
Although growth factors limit the growth and survival of individual cells by controlling their access to nutrients, other systems regulate nutrient transporter expression with the goal of meeting the nutritional needs of the whole organism. Multiple organ systems, such as the central nervous system, liver, kidney, pancreas and intestine, participate in homoeostatic feedback loops that work to maintain constant levels of nutrients in the blood-stream in part by altering nutrient transporter expression on the surface of their constituent cells. These control circuits that maintain organismal nutrient homoeostasis are quite distinct from those that produce tissue homoeostasis by controlling cell size and numbers that are the subject of this review. Although the signalling pathways that control nutrient transporter expression in response to fluctuating nutrient levels may provide mechanistic insight into how cell growth and survival signals affect nutrient transporter expression, the extensive literature on this subject is beyond the scope of this review. Instead, this review will cover the literature supporting the relatively recent idea that growth factors regulate cell growth and survival by restricting the ability of an individual cell to meet its own metabolic demand for nutrients. The implications of these results for the study of human disease will also be discussed.
WHAT ARE ‘NUTRIENT TRANSPORTERS’?
‘Nutrient’ is a very general term encompassing both molecules that are used as building blocks and those that are utilized in the cellular processes that generate ATP. Sugars, amino acids, metals, lipids and vitamins are all examples of nutrients that are brought into the cell using specific transporter proteins. The ‘transporters’ for these various nutrients are present in two main types: (i) true nutrient transporters (e.g. sugar and amino acid transporters) that form a pore or a channel to move specific nutrients across the cell membrane, and (ii) nutrient receptors, molecules such as the TfR (transferrin receptor) or the LDL (low-density lipoprotein) receptor that deliver their cargo via endocytosis of the nutrient-bound receptor. This functional difference has important implications for how the activity of transporters of these two classes might be regulated. Increasing the rate of endocytosis of the first class of transporters would decrease nutrient uptake, whereas the same change for the second class might increase nutrient acquisition as endocytosis is part of their nutrient uptake mechanism. As will be described, the expression of both types of transporters is positively regulated by growth signals.
EXTRINSIC SIGNALS REGULATE NUTRIENT TRANSPORTER EXPRESSION
Growth factors regulate both the production and destruction of nutrient transporter proteins (Figure 1). Many growth signals increase the transcription of nutrient transporter mRNAs, and microarray technology has made it relatively simple to detect changes in mRNA levels for a variety of transporters simultaneously. In contrast, much less is known about how signal transduction cascades affect the trafficking and degradation of these proteins, although this issue is beginning to be addressed [2–6]. Despite the fact that GLUT4 (glucose transporter 4) trafficking in response to insulin signalling has been very well studied because of the implications for diabetes mellitus, many of the molecular details are still lacking (reviewed in ). Furthermore, GLUT4 trafficking patterns may be very different from those of other nutrient transporter proteins. GLUT4 is expressed only in muscle cells and adipocytes in response to insulin signalling, and it is not yet clear whether stable intracellular pools of other transporters exist. Although the endocytosis and trafficking of the TfR and the LDL receptor has also been extremely well characterized, these ‘constitutively recycling’ proteins may be trafficked differently when growth factors become limiting. As the trafficking of nutrient transporter proteins has important implications for cell growth and survival , additional work in this area is essential.
Although the mechanistic details are not yet completely understood, the PI3K (phosphoinositide 3-kinase) pathway has emerged as a key regulator of both nutrient transporter expression and trafficking in response to extrinsic signals. This function is entirely consistent with the established role of this signalling pathway in promoting cell growth and survival. Recent reviews have summarized our current understanding of the PI3K pathway [9,10], but a brief overview of the most relevant aspects is presented in Figure 2. PI3K is a lipid kinase that produces the phosphoinositol phosphates PtdIns(3,4)P2 and PtdIns(3,4,5)P3. These molecules bind to the PH (pleckstrin homology) domain of Akt (also called protein kinase B), recruiting Akt to the plasma membrane, where it becomes phosphorylated and activated. Akt indirectly activates mTOR (mammalian target of rapamycin) by phosphorylating TSC (tuberous sclerosis complex) 2, thereby decreasing the ability of the TSC1/TSC2 complex to serve as a GAP (GTPase-activating protein) for the Rheb GTPase that binds to and activates mTOR. Akt-mediated phosphorylation of PRAS40 may also activate mTOR .
It is important to note that mTOR is found in two distinct complexes, mTORC (mTOR complex) 1 and mTORC2 (reviewed in [9,10]). mTORC1 is inhibited by the drug rapamycin and is responsive to Rheb GTP-binding state and to amino acid levels via the class 3 PI3K, Vps34. mTORC2, in contrast, is not inhibited by rapamycin and it remains unclear which, if any, upstream signals feed into this complex. mTORC1 and mTORC2 also have distinct substrates: mTORC1 stimulates translation by phosphorylating p70S6K (p70 S6 kinase) and 4EBP1 [eukaryotic initiation factor 4E (eIF4E)-binding protein 1], whereas mTORC2 phosphorylates the activation loop of Akt. Thus mTORC2 has the capacity to increase mTORC1 activity through its effects on Akt. A further wrinkle is that, although mTORC2 is insensitive to rapamycin treatment, long-term incubation of some cell types, particularly haemopoetic cells, with the drug does lead to mTORC2 inhibition. This presumably occurs because a shared component of mTORC1 and mTORC2 becomes limiting as it is held in inactive mTORC1–rapamycin complexes. At present, only a few direct mTOR substrates have been identified, and these seem insufficient to account for the diverse activities of this kinase. Thus, although much is known about signalling through the PI3K pathway, new mTOR substrates are likely to be identified.
Evidence has been rapidly accumulating that, in addition to its role in insulin-dependent glucose uptake, the PI3K pathway promotes the uptake of a wide variety of nutrients downstream of many different growth factor receptors (Table 1). Although the expression of transporters for many additional nutrients may be similarly regulated, to date, the expression of only four types of transporters has been shown to be regulated by extrinsic cell growth and survival signals.
Although GLUT4 has received a great deal of attention, much less is known about the regulation of the ubiquitously expressed glucose transporter, GLUT1, by growth signals. However, studies in lymphocytes clearly link GLUT1 expression to cell growth and proliferation. T-cell receptor stimulation increases GLUT1 mRNA and protein levels as long as the co-stimulatory molecule CD28 is also engaged [12,13]. Interestingly, antibody-mediated cross-linking of CTLA-4 (a negative regulator of T-cell activation ) prevents the CD28-dependent increase in GLUT1 expression . Thus limiting GLUT1 expression may be an important mechanism for constraining T-cell growth. Similarly, engagement of FcγRIIB, which inhibits B-cell growth, prevents GLUT1 up-regulation following stimulation through the B-cell receptor . Although PI3K activation is required in both T- and B-cells for GLUT1 up-regulation, T-cells, but not B-cells, depend upon mTOR signalling for this effect . Notch activation also up-regulates GLUT1 expression, probably by activating Akt . Notch provides essential survival and growth signals to pre-T-cells independently of the T-cell receptor.
Many cytokines that promote lymphocyte growth and proliferation also increase GLUT1 expression. IL (interleukin)-2 stimulation in the presence of PHA (phytohaemagglutin) up-regulates GLUT1 in human T-lymphocytes . TGF-β (transforming growth factor β) treatment increases GLUT1 mRNA and protein levels in freshly isolated quiescent cord blood CD4+ lymphocytes . IL-7, a cytokine essential for T-cell development, promotes GLUT1 expression through a PI3K-dependent mechanism . IL-3 stimulates GLUT1 surface expression in a PI3K-dependent manner . IL-3 transcriptionally up-regulates GLUT1, but also has post-translational effects on trafficking that are likely to be mediated by Akt and mTOR ([2,6,13,21,22] and A. L. Edinger, unpublished work). IL-3 withdrawal increases the rate at which GLUT1 is internalized, and this effect can be blocked by constitutively active Akt . In contrast with results obtained with other nutrient transporters , the ability of Akt to support GLUT1 surface expression appears to be rapamycin-insensitive . Akt may regulate GLUT1 expression through its ability to inhibit GSK3 (glycogen synthase kinase 3), as constitutively activating GSK3 decreases surface levels of GLUT1 . Interestingly, cytokine levels appear to function as a rheostat rather than an on/off switch: stepwise increases in the amount of IL-3 in the medium produce matching increases in GLUT1 expression .
Less information is available about the regulation of GLUT1 expression by growth and survival factors in non-haemopoetic tissues or by PI3K-independent signalling. Treatment of L6 myotubes with insulin results in an increase in GLUT1 mRNA and protein levels that is blocked by rapamycin [24,25]. Insulin also stimulates GLUT1 transcription in mouse hepatoma cells through an Akt-dependent mechanism . Similarly, treatment of mice with Akt-dependent prostate intraepithelial neoplasia with rapamycin decreased GLUT1 mRNA and protein expression . Thus the PI3K pathway also controls GLUT1 expression outside of the immune system. Several kinases that may feed into the PI3K pathway have also been shown to regulate GLUT1 expression. SGK1 (serum- and glucocorticoid-induced protein kinase 1), a serine/threonine kinase that is activated by PI3K and is highly homologous with Akt, promotes GLUT1 expression in Xenopus oocytes and transfected mammalian cells, but a role for endogenous SGK regulating GLUT1 expression in response to normal physiological signals has not been established . BCR (breakpoint cluster region)–Abl, the fusion protein responsible for most chronic myelogenous leukaemias, supports growth-factor-independent GLUT1 expression by affecting both its transcription and trafficking . v-Abl also promotes glucose uptake . The up-regulation of GLUT1 by BCR–Abl is blocked by the Abl kinase inhibitor imatinib (also known as Gleevec), suggesting that nutrient transporter down-regulation may contribute to the ability of this chemotherapeutic drug to kill tumour cells . As the PI3K inhibitors wortmannin and LY294002 also inhibit BCR–Abl-dependent GLUT1 expression, signalling through Akt and possibly mTOR is likely to be involved. Src [30,31] and Ras  also up-regulate GLUT1 transcription, although whether this translates into increased cell-surface levels of GLUT1 protein was not evaluated. The importance of PI3K/mTOR signalling downstream from Ras is also unclear.
Amino acid transporters
Mammalian cells express a wide variety of amino acid transporters with different transport specificities (for a review of the mammalian amino acid transport systems, see ). However, many of these have only recently been molecularly cloned, and few antibodies recognizing amino acid transporters have been described. This may help explain why the cell-surface expression pattern of most amino acid transporters is rather poorly characterized. The 4F2hc (4F2 heavy chain, also known as CD98) is one of the best-studied amino acid transporters, perhaps because it was cloned over 25 years ago and because multiple antibodies recognizing this protein have been developed. 4F2hc is a type II membrane protein and is not itself an amino acid transporter . Instead, 4F2hc forms disulfide-linked heterodimers with a variety of multiple membrane-spanning light chains [e.g. LAT1 (L-type amino acid transporter 1)] that are responsible for the amino acid transport properties of the complex. Another nearly ubiquitous murine cationic amino acid transporter, mCAT1, has also been cloned for a number of years and is similarly well studied . To date, most information regarding the regulation of amino acid transporter expression by growth signals comes from studies of these two transporters of both essential and non-essential amino acids.
As with glucose transporters, the effect of growth and proliferative signals on amino acid transporter expression has been best characterized in the immune system. T-cell activation increases 4F2hc expression prior to entry into the cell cycle . Along with 4F2hc, the 4F2 light chain LAT1 is up-regulated at the mRNA and protein levels in Jurkat T-cells by stimulation with PMA and ionomycin . The proliferation-inducing cytokines IL-2 and, to a lesser extent, IL-15, also increase 4F2hc mRNA levels in T-cells [37,38]. Stimulation of T-cells with concanavalin A or of B-cells with LPS (lipopolysaccharide) increases mCAT1 mRNA levels , and mCAT2 gene expression is induced by T-cell activation . IL-3 is required for 4F2hc and mCAT1 cell-surface expression in the murine haemopoetic cell line, FL5.12 ([4,5,8] and A. L. Edinger, unpublished work). Outside of the immune system, insulin treatment increases the levels of mCAT1 mRNA in quiescent rat liver  and, although it is unclear whether the increase in mRNA translates into increased protein expression in all cases, angiotensin II, PDGF (platelet-derived growth factor) and TNFα (tumour necrosis factor α) increase mCAT1 mRNA levels in various cell types (reviewed in ). PDGF treatment can also increase the levels of LAT1 mRNA . Thus a variety of extrinsic growth signals up-regulate amino acid transporter expression.
The PI3K pathway also plays a critical role in boosting amino acid transporter expression in response to growth signals. For example, IL-2-dependent up-regulation of 4F2hc is abrogated by treatment with the PI3K inhibitor LY294002 . Rapamycin, the specific mTOR inhibitor that inhibits T-cell proliferation, blocks 4F2hc up-regulation in stimulated T-cells . Treatment of human BJAB B-lymphoma cells with rapamycin decreases the mRNA levels of at least five different amino acid transporters . Furthermore, the expression of activated mutants of either Akt or mTOR is sufficient to support the expression of 4F2hc in growth-factor-deprived cells [4,5]. Although these findings suggest that mTORC1 modulates amino acid transporter expression, the observation that a kinase-dead mutant of mTOR, but not rapamycin treatment, alters 4F2hc subcellular localization in the IL-3-dependent FL5.12 cell line suggests that mTORC2 might also affect amino acid transporter trafficking, perhaps through effects on the actin cytoskeleton [3,44,45]. Apart from the PI3K pathway, few signal transduction molecules affecting amino acid transporter expression have been identified. PKC (protein kinase C) activation, however, reduces surface levels of mCAT1 in a glioblastoma cell line .
The transferrin receptor
Iron, an important co-factor for growth-promoting enzymes (e.g. ribonucleotide reductase) and for energy generation in the mitochondria, is brought into the cell via the TfR (reviewed in ). TfR expression is regulated by intracellular iron levels and by hypoxia (reviewed in [48–50]), but much less is known about how growth signals affect TfR expression. As with glucose and amino acid transporters, TfR expression is up-regulated by lymphocyte activation. The TfR (also known as CD71) is undetectable on resting lymphocytes, but its expression is dramatically increased on both B- and T-cells following antigen receptor ligation [51,52], by LPS in B-cells  and in a PI3K-dependent manner by IL-7, IL-2 and, to a lesser extent, IL-15 in T-cells [19,37,38]. The level of expression of the TfR in T-cells correlates well with proliferation as both rapamycin-treated and anergic T-cells fail to up-regulate the TfR following stimulation . mTOR also regulates TfR expression in response to IL-3 signalling [4,6,55]. In addition, transforming mutations in Akt prevent TfR down-regulation following growth factor withdrawal  and the TfR is a direct transcriptional target of the proto-oncogene c-Myc .
The LDL receptor
The synthesis of new cellular membranes requires cholesterol. In an actively dividing cell, it is likely that extracellular sources of cholesterol (LDL particles in the blood) are an important supplement to that produced within the cell by HMG-CoA (3-hydroxy-3-methylglutaryl-CoA) reductase. Thus it stands to reason that LDL receptor expression would also be positively regulated by growth signals. Although little information about how LDL receptor expression might be modulated by growth factors is available at this time, IL-3 stimulation supports LDL receptor expression probably through the activation of Akt and mTOR .
The role of the PI3K signal transduction pathway as a master regulator of nutrient transporter expression in mammalian cells is entirely consistent with how permease expression is regulated in model organisms. In yeast, the expression of a panoply of nutrient transporter proteins is controlled by TOR (Table 2). Rheb also regulates arginine and lysine uptake in yeast [57,58]. In Drosophila, the cell-surface expression of Slimfast, a cationic amino acid transporter, is positively regulated by several components of the PI3K pathway: TSC1/TSC2, Rheb and TOR [59,60]. This conservation of the ability of the PI3K pathway to promote nutrient transporter expression from yeast to mammals is consistent with the idea that ancient cell-autonomous signalling pathways have been re-wired to growth factor receptors as a mechanism for maintaining tissue homoeostasis.
THE MOLECULAR DETAILS OF NUTRIENT TRANSPORTER REGULATION: LESSONS FROM YEAST
While the upstream signal transduction cascades that regulate mammalian nutrient transporter expression are coming to light, virtually nothing is known about the downstream molecular mechanisms behind these effects. What transcription factors are involved? What are the post-translational signals that control nutrient transporter trafficking? Yeast may provide some clues. Many of the yeast transporters listed in Table 2 are regulated by TOR at the transcriptional level. In some cases, this involves TOR-mediated repression of the GATA transcription factors GLN3 and GAT1 . GATA-1 [70,71] and GATA-2  up-regulate TfR expression in erythroid cells and it will be interesting to determine whether GATA transcription factors affect the transcription of other mammalian nutrient transporter genes. On the other hand, ubiquitination is a post-translational modification that plays an important role in controlling nutrient transporter trafficking in yeast [73–75]. For example, the inverse regulation of the yeast tryptophan transporter TAT2 and the general amino acid permease GAP1 in response to TOR signalling reflects differences in the ubiquitination of these proteins (Figure 3, reviewed in ). GAP1 is expressed on the cell surface only when yeast are grown on a suboptimal nitrogen source such as urea. On preferred nitrogen sources, GAP1 ubiquitination results in its degradation in the vacuole (the yeast equivalent of the lysosome). In contrast, TAT2 is found on the plasma membrane only when yeast are grown on a high-quality nitrogen source such as ammonium. When conditions deteriorate, TAT2 is ubiquitinated and is trafficked to the vacuole for degradation. The expression of other yeast transporters is also regulated by ubiquitination, but it is not clear that this is in response to changes in TOR signalling in all cases (reviewed in ).
For TAT2 and GAP1, some of the proteins involved in these ubiquitin-dependent trafficking reactions have been defined. TOR modulates the phosphorylation status and probably the activity of the NPR1 kinase that in turn regulates the ubiquitination of GAP1 and TAT2 by RSP5 [63,65,78–80]. More recently, a complex named GSE (GTPase-containing complex for Gap1p sorting in the endosomes) was shown to regulate GAP1 sorting in response to ubiquitination and is required for GAP1 expression on the plasma membrane . Interestingly, three components of the GSE complex are also found in the EGO complex that is required for exit from rapamycin-induced growth arrest . Whether the EGO complex might control GAP1 trafficking following rapamycin removal was not examined. What do these yeast studies tell us about mammalian cell biology? Although several components of the GSE complex do not have clear mammalian homologues, the GTR1/GTR2 GTPases do . In fact, the mammalian GTR1 homologue, RagA, can substitute for the yeast protein, suggesting that a GSE-like complex might form in mammalian cells . Although mTOR clearly regulates amino acid transporter expression, the NPR1 kinase that functions downstream from yeast TOR does not have a clear mammalian homologue. Thus a more divergent kinase may fulfil a similar role, or mTOR might regulate transporter expression more directly. Interestingly, the RSP5 homologue Nedd4 ubiquitinates and down-regulates the epithelial sodium channel ENaC . Furthermore, activation of the Akt-related kinase SGK by aldosterone prevents Nedd4–ENaC binding, thereby increasing the cell-surface expression of this sodium channel. Whether Nedd4 is a true orthologue and regulates nutrient transporter expression remains to be determined.
Although the expression of many yeast permeases and mammalian cell-surface proteins such as ENaC, chemokine receptor CXCR4 and the EGF (epidermal growth factor) receptor is regulated by ubiquitination [77,84–86], ubiquitination of mammalian nutrient transporter proteins has not yet been described. However, modification of GLUT1 and GLUT4 by the ubiquitin-related molecule SUMO (small ubiquitin-related modifier) has been reported to decrease GLUT1 expression and increase GLUT4 expression . The importance of this finding for their regulation by trophic and survival signals remains unclear. Although a TfR–ubiquitin fusion protein is trafficked to the lysosome, most of these molecules are not degraded, suggesting that other signals (or additional ubiquitins) may be required for efficient turnover [88,89]. Apart from ubiquitin modification of the transporters themselves, ubiquitin modification of endocytic regulatory proteins, such as Eps15 and Hrs, may also play an important role in directing nutrient transporter trafficking in response to growth factor signalling . It will also be interesting to determine whether transporter phosphorylation affects their turnover rate as has been observed for plasma membrane proteins in yeast (reviewed in ). In summary, while studies in both yeast and mammalian cells suggest that ubiquitination may be important, there is a pressing need for studies addressing how nutrient transporter proteins are targeted for destruction and trafficked to the lysosome in mammalian cells.
EVIDENCE THAT CHANGES IN NUTRIENT TRANSPORTER EXPRESSION CONTROL CELL GROWTH AND SURVIVAL
Growth factor deprivation eliminates nutrient transporter expression (Figure 1). Might the loss of nutrient access be the stimulus that initiates growth-factor-withdrawal-induced apoptosis? The metabolic changes that accompany nutrient transporter down-regulation could lead to the activation of pro-apoptotic molecules such as Bak and Bax, committing the cell to death. On the other hand, rather than causing the cell's demise, the decline in nutrient transporter expression may simply correlate with death, reflecting a decreased metabolic demand in the dying cell (Figure 4). One way to discriminate between these competing models would be to maintain nutrient transporter expression in a cell deprived of growth factors. If dying cells simply require fewer nutrients, then maintaining nutrient transporter proteins should have no effect on cell viability. In contrast, if loss of nutrient access causes a metabolic catastrophe that initiates apoptosis, maintaining nutrient transporter expression should prevent apoptosis.
One way to produce growth-factor-independent nutrient transporter expression is to constitutively activate components of the PI3K pathway [4–6,16,21]. However, PI3K signalling promotes cell survival through multiple distinct mechanisms, and the relative importance of maintaining nutrient transporter expression for cell survival is difficult to gauge. An alternative approach is to overexpress nutrient transporters with the hope of saturating the machinery responsible for their down-regulation, thereby prolonging cellular access to nutrients. In fact, GLUT1 overexpression delays apoptosis in response to growth factor withdrawal , and co-expression of hexokinase to trap glucose inside the cell further increases survival . Similarly, stable expression of the amino acid transporter-associated protein 4F2hc in murine NIH 3T3 fibroblasts facilitates tumour formation in nude mice [93,94]. However, as 4F2hc is also involved in integrin signalling, its transforming ability might be also be related to this function . Furthermore, overexpression of the TfR promotes cellular proliferation under low-serum conditions and tumour formation in nude mice . Of course, a significant limitation of these overexpression studies is that uptake of only a single nutrient is maintained. Although some nutrients may play a more important role in supporting cell growth and survival than others, a diverse array of transporters are down-regulated by growth factor withdrawal, and a deficiency of even one critical nutrient would probably be sufficient to cause cell death. Thus a strategy that maintains global nutrient transporter expression would be a better test of whether nutrient transporter expression is sufficient to confer robust growth-factor-independent cell survival.
One approach to global nutrient transporter maintenance is to inactivate Rab7. Rab7 is a small GTPase that facilitates fusion events between late endosomes and lysosomes [96–98]. Among the proteins that regulate nutrient transporter endocytosis and lysosomal trafficking, Rab7 is an ideal protein to target as blocking its function results in endocytosed proteins being recycled back to the cell surface rather than leaving them stuck in endosomes (Figure 5). When Rab7 is inactivated using a dominant-negative mutant or by RNAi (RNA interference), down-regulated nutrient transporter proteins are re-expressed on the cell surface, and cells are largely protected from growth-factor-withdrawal-induced apoptosis . Consistent with this, blocking Rab7 function also facilitates cellular transformation, indicating that Rab7 might function as a tumour-suppressor gene. It is an important point that growth-factor-receptor-mediated signal transduction is not maintained in these cells; Akt activity, for example, is not maintained by inhibiting Rab7. Thus cell survival due to blocking Rab7 function is distinct from that produced by the loss of tumour-suppressor genes such as Cbl that prolong growth factor signalling by blocking receptor degradation . In summary, these results demonstrate that changes in nutrient transporter trafficking leading to the loss of nutrient access play a critical role in the initiation of growth-factor-withdrawal-induced apoptosis.
To confirm the results obtained in cells with reduced Rab7 activity, it will be important to develop independent approaches to maintaining nutrient transporter protein expression in the face of growth factor deprivation. Simply blocking endocytosis at an earlier step is unlikely to be as effective as inhibiting Rab7. Blocking global endocytosis is rapidly toxic to cells and would prevent the uptake of iron and LDL particles. On the other hand, mutations that disrupt trafficking to the lysosome, but trap nutrient transporters in the endocytic pathway would not increase nutrient uptake; intracellular transporters do not do the cell any good. Given these considerations, an increased understanding of how nutrient transporter endocytosis and trafficking is regulated will be essential to develop experiments to explore this model further. Furthermore, it should be taken into account when designing these experiments that cell lines are often derived from tumours or from cells that have become partially or completely transformed while adapting to in vitro growth conditions. As a result, these cells may carry mutations in the pathways that control nutrient transporter turnover. For example, PTEN (phosphatase and tensin homologue deleted on chromosome 10) deletion is very common in human tumours. Loss of PTEN constitutively activates the PI3K pathway which is likely to affect nutrient transporter expression and trafficking. Studies in primary cells or cells that are immortalized but not transformed (such as the IL-3-dependent cells used in the Rab7 studies [100,101]) will be necessary to help ensure that results obtained in cell lines accurately reflect how nutrient transporter proteins are trafficked in normal mammalian cells in situ.
The immune system offers additional examples of regulated nutrient access being used as a mechanism for tissue homoeostasis. IDO (indoleamine-2,3-dioxygenase), an enzyme produced by macrophages, suppresses effector T-cell function by catabolizing the essential amino acid tryptophan (reviewed in ). Cancer cells and antigen-presenting dendritic cells in tumour-draining lymph nodes also express IDO, facilitating immune escape by tumours (reviewed in [103,104]). IDO expression may also have an impact on nutrient transporter expression: tumours may be forced to adapt to a low tryptophan environment by up-regulating tryptophan transporters as they balance their own metabolic needs against the benefits of immune suppression. Another example of nutrient access limiting cell growth is the inability of lymphocytes to take up cystine, the oxidized form of cysteine that predominates in the oxidizing extracellular environment. Cysteine is rate-limiting for T-cell growth after stimulation, and antigen-presenting cells positively regulate T-cell growth and proliferation by making the reduced form of this amino acid available to lymphocytes (reviewed in ). These examples help to demonstrate the feasibility of limiting cell growth and survival by controlling nutrient access.
IMPLICATIONS OF REGULATED NUTRIENT TRANSPORTER TURNOVER FOR HUMAN DISEASE
Given the observations that cellular transformation is facilitated both by overexpressing nutrient transporters [22,56,92–94] and by blocking their degradation [4,5,8], nutrient transporter down-regulation may be an important mechanism for tumour suppression . Consistent with this idea, it is well documented that many human tumour cells express high levels of glucose transporters, amino acid transporters and the TfR [107–109]. Furthermore, the level at which these transporters are expressed often correlates with the degree of malignancy and has prognostic value. More malignant salivary gland tumours express higher levels of GLUT1, and, consistent with this, high GLUT1 levels correlate with lower patient survival rates . Similarly, elevated expression of the 4F2 light chain, LAT1, is associated with more advanced glioblastomas and a poorer prognosis . High levels of 4F2hc in human primary breast cancers also correlate with low patient survival . This study also reported that LDL receptor mRNA levels were elevated in tumours relative to normal breast tissue. Likewise, high-grade B-cell lymphomas express more TfR molecules than low-grade cancers . In fact, many tumour imaging approaches rely on the fact that tumour cells express high levels of nutrient transporters to distinguish cancerous tissue from the surrounding normal cells. FDG (2-[18F]fluoro-2-deoxy-D-glucose)-PET (positron emission tomography) scans detect tumours based on their increased uptake of FDG . Radiolabelled amino acids have also been used for tumour imaging . PET scan intensity (which probably reflects nutrient uptake capacity) generally correlates with tumour grade. Increased uptake of the tracer can also predict increased metastatic potential, reduced response to therapy and decreased patient survival. The increased efficiency with which cancer cells steal nutrients from the blood supply due to their elevated transporter expression levels may also contribute to the muscle wasting and cachexia seen in cancer patients.
If tumour cells exhibit high rates of nutrient uptake, could limiting nutrient transporter expression be an effective approach to cancer therapy? This strategy would be the converse to angiogenesis inhibitors: rather than limiting the delivery of necessary molecules to the tumour, such therapies would limit a tumour cell's access to the available nutrients. It has long been known that anti-TfR antibodies can inhibit cell growth [51,115], and these antibodies also exhibit anti-tumour activity . Reducing TfR levels by RNAi similarly results in cell cycle arrest . Consistent with these studies, iron chelators have shown some efficacy as cancer chemotherapeutics, and newer derivatives with higher oral bioavailability and longer half-lives are currently in clinical trials . Antibodies against 4F2hc also inhibit the proliferation of a variety of tumour cell lines , although this may be due to disruption of the interaction of 4F2hc with integrins rather effects on amino acid import . Reducing GLUT1 expression also results in decreased tumour cell proliferation, colony formation in soft agar, tumour growth and tumour invasiveness (reviewed in ). Furthermore, drugs that control nutrient availability in the extracellular space (e.g. asparaginase and arginine deiminase [119,120]) can limit the growth of several types of cancer, suggesting that therapies that directly affect nutrient transporter expression may be similarly effective. Given that these single-transporter approaches have shown promise, therapies designed to restrict global transporter expression are likely to be even more effective. A better understanding of nutrient transporter trafficking would greatly facilitate the development of such broadly acting drugs.
The role of transporter turnover in cell growth and survival also has important ramifications for the study of the role of autophagy in human disease. Interest in autophagy has grown rapidly over the last several years as more and more studies link this process to cellular transformation and cancer (see, for example, [121–128]) as well as to neurological diseases (reviewed in [129,130]). Through autophagy (literally, eating oneself), cells degrade their own components to produce energy during periods of nutrient stress . Damaged organelles are also turned over through this mechanism. The induction of autophagy in growth-factor-deprived cells appears to be an adaptive response to the loss of nutrient transporter expression. Consistent with this, autophagy is required for survival following growth factor withdrawal [124,132]. This has implications for oncogenesis, where a cell on the path to full transformation may have short-circuited apoptosis, but not yet acquired a mutation that will support growth-factor-independent nutrient uptake and will thus be dependent on autophagy. On the other hand, multiple studies have found that autophagy can also limit tumorigenesis (reviewed in ). Autophagy may suppress transformation by eliminating damaged organelles that produce DNA-damaging free radicals or through an incompletely understood autophagic cell death pathway [128,134]. Significantly, one thing missing from most of the studies linking autophagy to human disease is an analysis of nutrient transporter expression. Knowledge of the cell's nutrient status would greatly facilitate interpretation of the function of autophagy under the experimental conditions: is autophagy an appropriate response to the loss of nutrient transporter proteins or an alternative mechanism for programmed cell death? Evaluating the levels of nutrient transporter proteins on the cell surface in autophagic cells would be one way to begin to address many of the unanswered questions surrounding how autophagy is triggered and what role it plays in the pathology of these diseases.
Regulated nutrient transporter expression also has implications for immune system diseases, viral infections and perhaps even aging. Given the importance of increased nutrient transporter expression for the growth and proliferation of stimulated lymphocytes, immune responses might be modulated therapeutically by drugs that affect nutrient transporter expression. For example, blocking nutrient transporter up-regulation might curtail an unwanted immune response and produce T-cell anergy [12,15,54]. Additional studies will be required to determine whether decreases in nutrient transporter expression contribute to activation-induced cell death in T-cells. As many viral receptors are nutrient transporter proteins (Table 3), viral infections may also affect nutrient uptake. For example, shed viral envelope proteins might negatively impact nutrient uptake in bystander cells. GLUT1 function is impaired by the surface subunit of the HTLV (human T-cell lymphotrophic virus)-1 and -2 retroviruses that use this protein as a cell-surface receptor [135,136]. As GLUT1 is highly expressed in the brain, shed envelope proteins might be responsible for some of the neuropathology associated with HTLV infection. Finally, given the relationship between caloric restriction and aging (reviewed in ), it will be interesting to determine whether reduced nutrient transporter expression might mimic the effects of decreased nutrient supply on lifespan.
The available evidence suggests that nutrient transporter proteins are not constitutively expressed housekeeping proteins, but instead play a primary role in regulating cell growth and survival in response to extrinsic signals. Despite the clear implications for human disease, there are large gaps in our knowledge of how nutrient transporter expression is regulated. Given the role that altered nutrient transporter trafficking appears to play in transformation, it will be particularly important to uncover the molecular signals (ubiquitination? phosphorylation?) that target nutrient transporter proteins for endocytosis and degradation. Such studies will allow the further exploration of the role of nutrient transporter turnover in oncogenesis and may suggest novel approaches to cancer chemotherapy. Many other key questions still require answers. Does altering nutrient transporter expression affect carcinogenesis or other diseases in animal models? What is the relationship between altered nutrient transporter expression and autophagy during oncogenesis? Do growth and survival factors also regulate differentiation by modulating nutrient transporter expression? Do all of these signals flow through the PI3K pathway? Loss of access to which nutrients is most critical in inducing cell death? Is nutrient transporter degradation a default pathway or are there signal transduction pathways that promote nutrient transporter destruction? What is the role of regulated nutrient transporter expression in non-bone-marrow-derived cells? The answers to these and many other open questions will have important implications both for our understanding of cellular growth control and survival and for human health and disease.
I am supported in part by a grant from the NCI (National Cancer Institute) of the National Institutes of Health, K08 CA100526.
Abbreviations: BCR, breakpoint cluster region; EGO, exit from rapamycin-induced growth arrest; 4EBP1, eukaryotic initiation factor 4E (eIF4E)-binding protein 1; 4F2hc, 4F2 heavy chain; ENaC, epithelial sodium channel; FDG, 2-[18F]fluoro-2-deoxy-D-glucose; GAP, GTPase-activating protein; GLUT, glucose transporter; GSE, GTPase-containing complex for Gap1p sorting in the endosomes; GSK3, glycogen synthase kinase 3; HTLV, human T-cell lymphotrophic virus; IDO, indoleamine-2,3-dioxygenase; IL, interleukin; LAT, L-type amino acid transporter; LDL, low-density lipoprotein; LPS, lipopolysaccharide; mCAT1, murine cationic amino acid transporter 1; mTOR, mammalian target of rapamycin; mTORC, mTOR complex; p70S6K, p70 S6 kinase; PDGF, platelet-derived growth factor; PET, positron emission tomography; PHA, phytohaemagglutinin; PI3K, phosphoinositide 3-kinase; PKC, protein kinase C; PTEN, phosphatase and tensin homologue deleted on chromosome 10; RNAi, RNA interference; SGK, serum- and glucocorticoid-induced protein kinase; TfR, transferrin receptor; TGF-β, transforming growth factor β; TNFα, tumour necrosis factor α; TSC, tuberous sclerosis complex
- © The Authors Journal compilation © 2007 Biochemical Society