암세포 아미노산(단백질 대사)에 대한 연구.. 2019년 nature

[문형철]

1. nature  
2. british journal of cancer  
3. review articles  
4. article

New aspects of amino acid metabolism in cancer

Download PDF

• Review Article
• Open access
• Published: 10 December 2019

Cancer Metabolism

New aspects of amino acid metabolism in cancer

• Lisa Vettore, 
• Rebecca L. Westbrook & 
• Daniel A. Tennant 

British Journal of Cancer volume 122, pages150–156 (2020)Cite this article

•  
•  
•  

Abstract

An abundant supply of amino acids is important for cancers to sustain their proliferative drive. Alongside their direct role as substrates for protein synthesis, they can have roles in energy generation, driving the synthesis of nucleosides and maintenance of cellular redox homoeostasis. As cancer cells exist within a complex and often nutrient-poor microenvironment, they sometimes exist as part of a metabolic community, forming relationships that can be both symbiotic and parasitic. Indeed, this is particularly evident in cancers that are auxotrophic for particular amino acids. This review discusses the stromal/cancer cell relationship, by using examples to illustrate a number of different ways in which cancer cells can rely on and contribute to their microenvironment – both as a stable network and in response to therapy. In addition, it examines situations when amino acid synthesis is driven through metabolic coupling to other reactions, and synthesis is in excess of the cancer cell’s proliferative demand. Finally, it highlights the understudied area of non-proteinogenic amino acids in cancer metabolism and their potential role.

암이 증식력을 유지하려면 

아미노산을 풍부하게 공급하는 것이 중요합니다. 

아미노산은 

단백질 합성을 위한 기질로서의 직접적인 역할 외에도

 에너지 생성, 

뉴클레오시드 합성 및 세포 산화 환원 항상성 유지에 중요한 역할을 할 수 있습니다. 

암세포는 

복잡하고 영양이 부족한 미세 환경 내에 존재하기 때문에 때때로 

대사 공동체의 일부로 존재하여 

공생 및 기생 관계를 형성하기도 합니다. 

실제로 이러한 현상은 

특정 아미노산에 대해 보조적인 역할을 하는 

암에서 특히 두드러집니다. 

이 리뷰에서는 

암세포가 안정적인 네트워크와 치료에 대한 반응으로 

미세 환경에 의존하고 기여할 수 있는 여러 가지 방법을 예시를 통해 설명함으로써

 기질/암세포 관계에 대해 논의합니다. 

또한 

아미노산 합성이 다른 반응과의 대사적 결합을 통해 추진되고 

합성이 암세포의 증식 수요를 초과하는 상황을 살펴봅니다. 

마지막으로, 

암 대사에서 비단백질 생성 아미노산에 대한 연구가 

부족한 영역과 그 잠재적 역할을 강조합니다.

Similar content being viewed by others

Amino acids in cancer

Article Open access24 January 2020

Amino acid metabolism in tumor biology and therapy

Article Open access13 January 2024

Tumour-reprogrammed stromal BCAT1 fuels branched-chain ketoacid dependency in stromal-rich PDAC tumours

Article 06 July 2020

Background

The major function of amino acids in mammalian cells is as substrates for new protein synthesis. There is therefore a significant demand for them in the proliferating cells within a tumour. Those that are essential – defined as those whose carbon skeletons cannot be synthesised by the cell, and include leucine, tryptophan and histidine – are required in significant amounts from both the diet and the intestinal microbiota. Non-essential amino acids can also be synthesised from endogenous sources, providing more flexibility for the cancer cell to ensure an adequate supply. While the definition of non-essential and essential is maintained in tumours at a system-wide level, this classification does not tell the whole story about how cancer cells balance supply and demand within a tumour. Studies over the past few decades have shed considerable light on how transport systems, stromal cells, gene silencing and redox homoeostasis can all play a role in how cancer cells maintain an adequate supply of amino acids to fulfil their proliferative drive. Similarly, they have also suggested novel means of targeting the supply of different amino acids to a cancer cell, with the potential for novel therapeutic interventions to block tumour growth in the future. This review will examine aspects of the control of amino acid metabolism, including when ‘non-essential’ amino acids become limiting for tumour growth, how the stroma is often utilised to provide these nutrients and conditions in which the synthetic pathways are regulated by other factors, sometimes producing amino acids in excess of tumour demand. Finally, we briefly discuss the emerging interest in non-proteinogenic amino acids in tumour metabolism.

배경

포유류 세포에서 

아미노산의 주요 기능은 

새로운 단백질 합성을 위한 기질입니다. 

따라서 

종양 내에서 증식하는 세포에는 

아미노산이 많이 필요합니다. 

필수 아미노산은 세포에서 탄소 골격을 합성할 수 없으며 

류신, 트립토판 및 히스티딘을 포함하는 아미노산으로 정의되며, 

식단과 장내 미생물 모두에서 상당한 양이 필요합니다. 

비필수 아미노산은 

내인성 공급원으로부터도 합성될 수 있으므로 

암세포가 적절한 공급을 보장할 수 있도록 

더 많은 유연성을 제공합니다. 

비필수 아미노산과 필수 아미노산의 정의는 

종양에서 시스템 전체 수준에서 유지되지만, 

이 분류는 암세포가 종양 내에서 공급과 수요의 균형을 유지하는 방법에 대한 전체 이야기를 말해주지는 않습니다. 

지난 수십 년 동안의 연구를 통해 

수송 시스템, 기질 세포, 유전자 침묵 및 산화 환원 항상성 등이 

암세포가 증식을 위해 적절한 아미노산 공급을 유지하는 데 

어떻게 작용하는지에 대한 상당한 사실이 밝혀졌습니다. 

마찬가지로, 

연구진은 암세포에 다양한 아미노산 공급을 표적으로 삼는 새로운 방법을 제안했으며, 

향후 종양 성장을 차단하는 새로운 치료적 개입의 가능성을 제시했습니다. 

이 리뷰에서는 

'비필수' 아미노산이 종양 성장에 제한이 되는 경우, 

기질이 이러한 영양소를 공급하기 위해 종종 활용되는 방법, 

합성 경로가 다른 요인에 의해 조절되어 

때로는 종양 수요를 초과하는 아미노산을 생산하는 조건을 포함하여 

아미노산 대사 제어의 측면을 검토합니다. 

마지막으로 

종양 대사에서 비단백질 생성 아미노산에 대한 

새로운 관심에 대해 간략하게 설명합니다.

Non-essential yet required?

A number of cancers have been found that are auxotrophic (i.e. depend on exogenous sources) for non-essential amino acids. In a number of cases this is through simple loss of expression‎ of an enzyme involved in its synthesis through direct mutation or silencing.1,2 Three such examples of this are the synthesis of arginine, asparagine and glutamine. In most cell types, arginine can be synthesised through the activity of argininosuccinate synthetase (ASS1) and argininosuccinate lyase (ASL), their combined activities transferring the amino group of aspartate to convert citrulline into arginine. It has been shown that ASS1 is not expressed in a number of different malignancies including some melanoma, prostate cancer, hepatocellular carcinoma, mesothelioma and bladder cancers.3,4,5 Given that this loss of ASS1 expression‎, and therefore the ability to synthesise arginine de novo is specific to malignant cells, trials of the non-mammalian enzyme, arginine deiminase (ADI), as a potential therapy to selectively kill the cancer cells have been performed and are ongoing.6,7

Silencing of asparagine synthetase (ASNS), the enzyme that uses the amide group from glutamine to synthesise asparagine from aspartate has also been shown as a cause of cancer-specific auxotrophy. It has been known since the 1970s that the response of paediatric acute lymphoblastic leukaemia (ALL) to therapy was inversely correlated with asparagine synthesis. Indeed, this observation led to the use of bacterial asparaginase (ASNase) to cure the majority of children with this cancer, either as single agent or as combination therapy,8 making it one of the most successful forms of metabolism-targeted therapy.

A further, less well-described auxotrophy is for glutamine, through loss or downregulation of glutamine synthetase (GS), which has been described in multiple myeloma, ovarian cancer and oligodendroglioma cells.9,10,11 GS synthesises glutamine from glutamate and NH4+, which has previously been shown to be important for continued tumour proliferation, particularly when glutamine may be limiting.12 The plasma membrane glutamine transporters may therefore represent potential therapeutic targets for tumours with GS deficiency. ASNase treatment, which depletes plasma glutamine in addition to asparagine, is also likely to show some efficacy in tumours lacking GS, although alternative means of depleting plasma glutamine by using phenylacetate have previously been trialled, albeit without stratification of patients for GS expression‎ status.13,14 Interestingly, stromal cell expression‎ of GS may also be important for cancer growth, suggesting that targeting this mechanism may indirectly reduce tumour cell proliferation.15

It is particularly interesting that these examples of tumour-associated auxotrophy are centred around two reactions that utilise aspartate to synthesise other amino acids, both directly (ASNS and ASS1) and indirectly (GS, through regeneration of glutamine required for ASNS activity). It is becoming clear that aspartate is a significant metabolic hub, and a major product of the oxidative glutaminolysis pathway. Aspartate is a required substrate for other anabolic pathways, including the synthesis of purines (both directly within the base carbon skeleton and amination of IMP to form AMP) and pyrimidines. It is an appealing hypothesis that aspartate may be limiting in some tumours,16 and that cancer cells selectively prioritise synthesis of nucleotides over asparagine and arginine, given that the amino acids may be more easily available within the microenvironment than substrates for the nucleotide salvage pathways. Indeed, this can be further highlighted by the recently reported depletion of aspartate observed in some rare tumour types driven by mutations in the metabolic enzyme, and a key component of the glutaminolytic pathway, succinate dehydrogenase.17,18

However, it is important to note that resistance to asparaginase and arginase therapy has been observed without re-expression‎ of ASS1 or ASNS. One potential short-term mechanism of resistance has been suggested as autophagy, by resupplying intracellular amino acids during therapies that deplete exogenous sources.19,20 However, this mechanism is not conducive to long-term survival in a proliferating tumour – a more dependable supply in the long term is likely to come from tumour-associated stromal cells.21 In the case of ASNase treatment of ALL, amino acid supply from mesenchymal-derived stromal cells (asparagine) and adipocytes (glutamine) has been reported, reducing the efficacy of therapeutic depletion of asparagine in the peripheral blood.21,22 This paradigm does not appear to end with those tumour cells that require specific amino acids for growth, but forms part of a continuum in which stromal and tumour cells swap nutrients including amino acids as part of a wider tumoural metabolic network.

필수적이지 않지만 필수적인 아미노산?

비필수 아미노산에 대한

보조적(즉, 외인성 공급원에 의존하는) 암이 다수 발견되었습니다.

대부분의 경우 이는

직접적인 돌연변이 또는 침묵을 통해 합성에 관여하는 효소의

단순한 발현 손실에 의한 것입니다.1,2

이러한 예로

아르기닌, 아스파라긴 및 글루타민 합성을 들 수 있습니다.

대부분의 세포 유형에서

아르기닌은

아르기니노숙시네이트 합성효소(ASS1)와

아르기니노숙시네이트 분해효소(ASL)의 활성을 통해 합성될 수 있으며, 이

두 효소의 결합 활동은 아스파르트산염의 아미노기를 전달하여

시트룰린을 아르기닌으로 전환합니다.

일부 흑색종, 전립선암, 간세포암, 중피종 및 방광암을 포함한

다양한 악성 종양에서 ASS1이 발현되지 않는 것으로 나타났습니다.3,4,5

이러한 ASS1 발현의 손실,

따라서 아르기닌 신규 합성 능력은 악성 세포에 특이적이기 때문에

암세포를 선택적으로 죽이는 잠재적 치료제로서

비포유동물 효소인 아르기닌 디이미니아제(ADI)의 시험이 수행되어 현재 진행 중입니다.6,7

글루타민의 아미드기를 사용하여

아스파르테이트에서 아스파라긴을 합성하는 효소인 아스파라긴 합성 효소(ASNS)의 침묵도

암 특이적 보조 영양증의 원인으로 밝혀졌습니다.

소아 급성 림프모구 백혈병(ALL)의 치료에 대한 반응이

아스파라긴 합성과 반비례한다는 사실은

1970년대부터 알려져 왔습니다.

실제로 이러한 관찰로 인해

박테리아 아스파라기나아제(ASNase)를 단일제 또는 병용 요법으로 사용하여

대부분의 소아암을 완치할 수 있게 되었고,8

이는 가장 성공적인 형태의 대사 표적 치료법 중 하나가 되었습니다.

다발성 골수종, 난소암 및 희소돌기아교종 세포에서

글루타민 합성효소(GS)의 손실 또는 조절 저하를 통한

글루타민에 대한 또 다른 잘 설명되지 않은 보조요법이 있습니다.9,10,11

GS는

글루타민산염과 NH4+로부터 글루타민을 합성하는데,

이는 특히 글루타민이 제한될 수 있는 경우

지속적인 종양 증식에 중요한 것으로 이전에 밝혀졌습니다.12

따라서

혈장막 글루타민 수송체는

GS 결핍 종양의 치료 표적이 될 가능성이 있습니다.

아스파라긴과 더불어 혈장 글루타민을 고갈시키는 ASNase 치료도

GS가 부족한 종양에서 어느 정도 효과를 보일 수 있지만,

이전에 페닐아세테이트를 사용하여 혈장 글루타민을 고갈시키는 대체 방법이 시도되었지만

환자의 GS 발현 상태를 계층화하지는 않았습니다.13,14

흥미롭게도

GS의 기질 세포 발현도 암 성장에 중요할 수 있으므로

이 메커니즘을 표적으로 삼으면 종양 세포 증식을 간접적으로 줄일 수 있음을 시사합니다.15

특히 이러한 종양 관련 보조증식의 예가 아스파테이트가 다른 아미노산을 합성하는 데 직접적으로(ASNS 및 ASS1) 그리고 간접적으로(GS, ASNS 활동에 필요한 글루타민 재생을 통해) 활용되는 두 가지 반응을 중심으로 이루어진다는 점이 흥미롭습니다. 아스파르트산염이 중요한 대사 허브이자 산화적 글루타민 분해 경로의 주요 산물이라는 사실이 점점 더 명확해지고 있습니다. 아스파르트산염은 퓨린(염기 탄소 골격 내에서 직접 합성 및 AMP를 형성하기 위한 IMP의 아미네이션)과 피리미딘의 합성을 포함한 다른 동화 경로에 필수적인 기질입니다. 일부 종양에서는 아스파르트산염이 제한적일 수 있으며,16 암세포가 뉴클레오타이드 회수 경로의 기질보다 미세 환경 내에서 아미노산을 더 쉽게 이용할 수 있다는 점을 고려할 때 아스파라긴과 아르기닌보다 선택적으로 뉴클레오타이드 합성을 우선시한다는 가설은 매력적인 가설이 될 수 있습니다. 실제로 최근 보고된 대사 효소의 돌연변이와 글루타민 분해 경로의 핵심 구성 요소인 숙시네이트 탈수소효소에 의해 일부 희귀 종양 유형에서 관찰된 아스파르트산염의 고갈은 이를 더욱 강조합니다.17,18

그러나 아스파라기나아제 및 아르기나제 치료에 대한 내성은 ASS1 또는 ASNS의 재 발현 없이 관찰되었다는 점에 유의하는 것이 중요합니다. 외인성 공급원을 고갈시키는 치료 중에 세포 내 아미노산을 재공급하는 자가포식이 내성의 한 가지 잠재적 단기 메커니즘으로 제시되었습니다.19,20 그러나 이 메커니즘은 증식하는 종양에서 장기 생존에 도움이 되지 않으며 장기적으로 더 신뢰할 수 있는 공급은 종양 관련 기질 세포에서 나올 가능성이 높습니다.21 ALL의 ASNase 치료의 경우, 중간엽 유래 기질 세포(아스파라긴)와 지방 세포(글루타민)에서 아미노산을 공급하여 말초 혈액에서 아스파라긴의 치료 효과를 감소시키는 것으로 보고되었습니다.21,22 이 패러다임은 성장을 위해 특정 아미노산이 필요한 종양 세포로 끝나지 않고 더 넓은 종양 대사 네트워크의 일부로서 기질 세포와 종양 세포가 아미노산을 포함한 영양소를 교환하는 연속체의 일부를 형성하는 것으로 보입니다.

The metabolic relationship between tumour and stroma: symbiotic or parasitic?

The trade in metabolites between stromal cells and the cancer cells themselves is becoming increasingly seen as an important facet of tumour metabolism, and often involves the transport of amino acids. One recently reported bidirectional relationship is between cancer-associated fibroblasts (CAFs) and cancer cells. In this model, it was shown that CAFs provide aspartate to cancer cells, which is taken up via the transporter SLC1A3 (also known as the excitatory amino acid transporter 1 [EAAT1]) to support nucleotide biosynthesis, while tumour cells reciprocate with glutamine-derived glutamate taken up by the CAFs through the same transporter.23 The nature of the transport(s) responsible for efflux of these two amino acids from cancer and stromal cells is not yet clear. Interestingly, the expression‎ of SLC1A3, a transporter originally described as being central in regulating extracellular glutamate levels in synapses, was more recently described as being regulated by p53 in cancer cells, providing the cells with aspartate during glutamine deprivation.24 This metabolic relationship has significant benefits for the cancer cell and may lead to less reliance on the oxidative TCA cycle for proliferation, given that glutamate is swapped for aspartate rather than oxidatively metabolised to synthesise it.25,26 This may have the effect of changing the profile of nutrients that are incorporated into the TCA cycle, reducing oxygen consumption by the cancer cells. The knock-on effect would be that the increased conversion of glutamate into aspartate by the stromal cells would probably have the opposite effect on them, i.e. by increasing their requirement to respire to oxidise NAD+ (see Fig. 2), thereby making them more sensitive to complex I inhibition, such as by metformin.

The amido group of glutamine is used to drive an entirely different spectrum of reactions compared with those utilising the amino group, including purine synthesis, asparagine synthesis and O-linked glycosylation (Fig. 1). The importance of the amido nitrogen to drive cancer cell proliferation has recently been described in glioma, where it was shown that glutamine-derived glutamate had two potential fates: amidation within the glioma cells by glutamine synthase (by using free ammonia) to regenerate glutamine, or export for uptake and amidation by astrocytes.12 This second system permits secretion of the glutamine by astrocytes and uptake by the glioma cells, effectively permitting the astrocytes to fix ammonia from the microenvironment to be utilised for nucleotide synthesis by the tumour.

종양과 기질 사이의 대사 관계: 공생 관계인가 기생 관계인가?

기질 세포와 암세포 사이의 대사 산물 교환은 종양 대사의 중요한 측면으로 점점 더 많이 인식되고 있으며, 종종 아미노산 수송과 관련되어 있습니다. 최근에 보고된 양방향 관계 중 하나는 암 관련 섬유아세포(CAF)와 암세포 사이의 관계입니다. 이 모델에서는 CAF가 암세포에 아스파르테이트를 제공하고, 이 아스파르테이트는 수송체 SLC1A3(흥분성 아미노산 수송체 1[EAAT1]이라고도 함)을 통해 흡수되어 뉴클레오티드 생합성을 지원하며, 종양 세포는 동일한 수송체를 통해 CAF가 흡수한 글루타민 유래 글루타메이트로 상호 교환하는 것으로 나타났습니다.23 암과 간질 세포에서 이 두 아미노산의 유출을 담당하는 수송체의 성격은 아직 명확하지 않은 것으로 밝혀졌습니다. 흥미롭게도, 원래 시냅스에서 세포 외 글루타메이트 수준을 조절하는 데 중심적인 역할을 하는 것으로 알려진 수송체 SLC1A3의 발현은 최근에 암세포에서 p53에 의해 조절되어 글루타민이 부족한 동안 세포에 아스파테이트가 공급되는 것으로 설명되었습니다.24 이러한 대사 관계는 암세포에 상당한 이점이 있으며, 글루타메이트가 산화적으로 대사되어 합성되는 것이 아니라 아스파테이트와 교환되기 때문에 증식을 위해 산화적 TCA 순환에 덜 의존하게 할 수 있습니다.25,26 이것은 TCA 순환에 통합되는 영양소의 프로필을 변경하여 암세포의 산소 소비를 줄이는 효과가있을 수 있습니다. 노크 효과는 기질 세포에 의한 글루타메이트의 아스파르테이트 전환이 증가하면 NAD+ 산화를 위한 호흡 요구량이 증가하여(그림 2 참조), 메트포르민과 같은 복합체 I 억제에 더 민감해지는 반대 효과를 가져올 수 있다는 것입니다.

글루타민의 아미도기는 퓨린 합성, 아스파라긴 합성 및 O-결합 글리코실화를 포함하여 아미노기를 사용하는 것과는 완전히 다른 스펙트럼의 반응을 유도하는 데 사용됩니다(그림 1). 암세포 증식을 촉진하는 아미도 질소의 중요성은 최근 신경교종에서 설명되었는데, 여기서 글루타민 유래 글루타메이트는 글루타민 합성효소(유리 암모니아 사용)에 의해 신경교종 세포 내에서 아미드화되어 글루타민을 재생하거나 성상세포의 흡수 및 아미드화를 위해 배출되는 두 가지 잠재적 운명을 가진 것으로 나타났습니다.12 이 두 번째 시스템은 성상교세포에 의한 글루타민 분비와 신경교종 세포에 의한 흡수를 허용하여 성상교세포가 미세 환경에서 암모니아를 고정하여 종양의 뉴클레오티드 합성에 활용할 수 있도록 효과적으로 허용합니다.

Fig. 1

Metabolism of glutamine. The amido group of glutamine is involved in relatively few reactions in addition to deamidation, some major examples are indicated. Transamination, for which only some representative reactions are shown, involves a number of 2-oxoacids that can be reversibly converted into the amino acid. Colours are used in both sets of reactions to indicate the enzyme (left) responsible for the reaction (right). On the left are the reactions that require, or evolve ammonia as part of the metabolism of glutamine to or from α-ketoglutarate. Abbreviations: αKG α-ketoglutarate, ALAT, alanine aminotransferase, ASNS, asparagine synthetase, BCAT, branched-chain aminotransferase, FGAM 5′-phosphoribosyl-N-formylglycinamidine, FGAR 5′-phosphoribosyl-N-formylglycinamide, FGARAT FGAR amidotransferase, GFAT glutamine fructose 6-phosphate amidotransferase, Gln glutamine, GLS glutaminase, Glu glutamate, GPAT glutamine phosphoribosyl pyrophosphate amidotransferase, GOT glutamic-oxaloacetic aminotransferase, GS glutamine synthetase, OAT ornithine aminotransferase, PRA 5′-phosphoribosyl-1-amine, PRPP 5′-phosphoribosyl-1-pyrophosphate, PSAT phosphoserine aminotransferase.

Full size image

The two examples highlighted above show two relationships between cancer and stromal cells: one in which both cells appear to benefit, and another in which the cancer cells take advantage of a physiologically relevant stromal activity. This is also the case when the relationship between the cell providing the amino acid and the cancer cell is physically more distant. For example, the process of tumour-associated cachexia, in which muscles and adipose tissue are progressively catabolised, is often observed in late-stage patients as their tumours continue to require nutrients in excess of what is contained in the diet.

An alternative means by which cancer cells can take up amino acids, but this time en masse, comes in the form of a process known as macropinocytosis. Described as ‘cell-drinking’, it is a process in which cells take up gulps of their microenvironment, known in a number of cell types under non-pathological conditions.27 Although first described in malignant cells in 1937,28 and then identified as being driven by oncogenic RAS in 1986,29 it was not until 2013 that a metabolic function was ascribed to macropinocytosis in cancer cells, when it was shown to result in uptake of whole proteins in pancreatic ductal adenocarcinoma (PDAC) cells.30 These proteins were catabolised through the autophagic machinery by the cancer cells, providing free amino acids for new protein synthesis, or catabolism to produce ATP.30,31 This process allows the cancer cells to survive in relatively nutrient-poor environments, taking up a number of different macromolecules derived from both the periphery (probably accumulating in oedema) and necrotic areas of the tumour. As well as being activated by other oncogenic signalling pathways, such as in those tumours driven by loss of Phosphatase and tensin homolog (PTEN) or aberrant WNT signalling,32,33 macropinocytosis can also be induced by the metabolic microenvironment of the tumour, as shown by macropinocytosis of albumin-associated free fatty acids by colorectal cancer cells through G protein-coupled receptor 120 (GPR120).34 Given that this process allows a tumour cell to continue to proliferate in the absence of abundant exogenous nutrients, it may play a particularly important role in hypoxic tumour regions, where not only are exogenous nutrients often limiting, but also where necrotic areas are juxtaposed to the malignant cells.

There is also evidence that cancers utilise amino acids derived from the extracellular matrix to support metabolism, in particular proline for ATP generation.35,36 In one study, the catabolism and uptake of collagen fragments were found to support PDAC tumour survival, by using macropinocytosis amongst other mechanisms when nutrient-deprived.36 Another study in breast cancer showed that in metastatic models of breast cancer, proline uptake and degradation supported survival.35 In this latter model, it is likely that plasma concentrations of proline can support this metabolic phenotype (~170 µM37), although digestion of the extracellular matrix (ECM) in the target tissue may also be important upon extravasation and colonisation in the absence of appropriate nutrient supply.

Tumours have therefore found means of utilising diverse mechanisms to provide themselves a supply of the amino acids that they require for malignant progression. It remains the case that many of these mechanisms are probable targets for novel therapy. However, given that a number of them are based upon non-pathological processes by which cells form a healthy metabolic community, care will need to be taken to avoid on-target side effects.

Metabolic coupling in the regulation of amino acid synthetic pathways

In common with most enzymatic reactions, non-essential amino acid synthesis requires more than one substrate and results in more than one product. Most often, this reaction is a simple transamination by using glutamate as the amino donor, resulting in the generation of α-ketoglutarate (Fig. 1). The direction of these reversible reactions depends on the local concentrations of the substrates and products as they are close to being energetically neutral. The relative concentrations of glutamate and α-ketoglutarate within the cytoplasm and mitochondria are therefore critical in determining which amino acid is made or catabolised, and how this reaction influences the wider metabolic network. A good example of this is the synthesis of aspartate. Within the mitochondrion, oxaloacetate (OAA) is most often a product of oxidative TCA cycle activity (Fig. 2), while glutamate is produced by mitochondrial glutaminase. Coupling these two reactions together in this compartment maintains flux of α-ketoglutarate into the TCA cycle, while removing OAA to permit continued activity, producing reducing potential to generate ATP.38 In this way, two metabolites central to proliferation – ATP and aspartate – are synthesised in parallel.25,26 These links go further, given that aspartate, glutamate, α-ketoglutarate and malate are functionally coupled through the malate–aspartate shuttle, which is important for moving reducing potential between the matrix and the cytosol (Fig. 2).

Fig. 2

Interplay between amino acid metabolism and redox homoeostasis. The synthesis and catabolism of amino acids is interwoven into the redox homoeostasis of the cell. The malate–aspartate shuttle, as well as moving NADH between the cytosol and the mitochondrial matrix, also moves the amino acids glutamate and aspartate between the two compartments, and is functionally connected to the TCA cycle. When aspartate is removed from this cycle to synthesise asparagine, arginine or nucleosides, this would disrupt the cycle, requiring additional carbon input. NADH oxidation reactions are shown in green, while NAD+ reduction is shown in red, indicative of the connectivity of this network. Amino acids are represented in blue, while proteins (transporters and electron carriers) are in orange. Finally, the subnetworks illustrated are outlined in blue (glycolysis), pink (electron transport chain), mauve (proline-redox shunt), green (TCA cycle) and cream (malate–aspartate shuttle). Abbreviations: αKG α-ketoglutarate, 1,3BPG 1,3-bisphosphoglycerate, 3PG 3-phosphoglycerate, 3PP 3-phosphopyruvate, AcCoA acetyl CoA, Ala alanine, Asn asparagine, Asp aspartate, Cit citrate, G3P glyceraldehyde 3-phosphate, Gln glutamine, Glu glutamate, Gly glycine, Lac lactate, Mal malate, NAD+ nicotinamide adenine dinucleotide, NADH reduced NAD+, OAA oxaloacetate, P5C pyrroline 5-carboxylate, Pro proline, Pyr pyruvate, Ser serine.

Full size image

Alanine is another amino acid, the synthesis of which (from pyruvate) is thought to lie mainly in the mitochondrial matrix.39 The alanine aminotransferase (ALAT) therefore competes for mitochondrial pyruvate with pyruvate dehydrogenase (PDH), which oxidatively decarboxylates to form acetyl-CoA (Fig. 2). It would therefore be expected that in conditions where PDH activity is lower, which includes hypoxia, increased alanine synthesis may be observed. However, recent evidence suggesting that ALAT activity is important for driving ECM formation through α-ketoglutarate production in metastatic breast cancer cells,40 has shed a different light on the role of amino acid synthesis. The data define this pathway as one that can be under regulation by coupled reactions – the synthesis of alanine is regulated under some conditions by a demand for α-ketoglutarate and driven by uptake of pyruvate from the microenvironment.40 Under such conditions, the fate of the alanine is not clear but is  probably in excess of what is required for anabolism – this would indicate that it is excreted into the extracellular space and would suggest the potential for it to be taken up by stromal cells, which may secrete the pyruvate required for the reaction themselves. Interestingly, this system demonstrates a reversal of that previously described in PDAC, where the major stromal cell type, the stellate cell, secretes alanine that is taken up by the cancer cells and deaminated to pyruvate as a significant carbon source.41 This means that while increasing oxygen consumption in what is a highly hypoxic tumour type – which appears somewhat counterintuitive – it also spares glucose to be used by other cell types or for synthesis of macromolecular substrates, including serine and glycine.

A further example of amino acid synthesis regulated by a coupled metabolic reaction was the recently described synthesis of proline in glioma. Our research group investigated the apparent drivers of the increased production and excretion of proline by cells expressing an oncogenic mutant form of the enzyme, isocitrate dehydrogenase 1 (IDH1), which is regularly observed in grade II and III gliomas and around 10% of grade IV gliomas (glioblastomas).42,43 The normal activity of IDH1 involves the NADPH-driven carboxylation of α-ketoglutarate to isocitrate, which is reversible depending on the redox state of the cell. However, the IDH1 mutation observed in gliomas results in a neomorphic activity that reduces α-ketoglutarate to 2-hydroxyglutarate, driven by the oxidation of NADPH.44 This has been suggested to perturb the redox homoeostasis of the cell.45,46,47 It has been previously well-documented that proline synthesis may be a means of responding to adverse environmental conditions, including redox stress, in plants.48 In mammalian systems, proline synthesis is through two routes: either in the cytosol from ornithine through the activity of ornithine aminotransferase and pyrroline 5-carboxylate reductase 3 (PYCR3, also known as PYCRL), or in the mitochondria from glutamate through either PYCR1 or 2. Each of these reductase reactions is linked to NADPH oxidation –PYCR3 appears to have a higher activity in the presence of NADPH, while PYCR1 and 2 favour NADH.49 Interestingly, we showed that cells expressing the oncogenic mutant of IDH1 increased proline synthesis and excretion, but this appeared to be specifically through the increased activity of mitochondrial PYCR1 by using glutamine as the major carbon source.42 The fact that proline synthesised in these conditions was excreted suggested that its synthesis may be driven by oxidation of NADH, rather than a demand for proline. Indeed, the reduction of glutamate to proline oxidises two moles of NADH per mole of proline synthesised, making it a very efficient means of responding to an increased demand for NAD+ (Fig. 2). Flux along this pathway is likely to be increased under general conditions in which the mitochondrial NADH:NAD+ ratio increases, which would include hypoxia and where there are deficiencies in NADH oxidation due to mitochondrial dysfunction. This mechanism is conceptually the same as reduction of cytosolic pyruvate by lactate dehydrogenase to regenerate NAD+ in the cytosol, although it may be more efficient in terms of NADH oxidised to carbons exported (2.5/NADH [proline] compared with 3/NADH [lactate]). However, the means of exporting proline from the matrix is currently unclear and this may add additional complexity to the system, as it is likely either to drive increased mitochondrial membrane potential (through proton symport) or require import of another metabolite into the matrix. Interestingly, there is a further potential role for increased proline synthesis in conditions where redox homoeostasis is perturbed. Proline can react with hydroxyl radicals to form moieties such as 4-hydroxyproline and 3-hydroxyproline, although it has been shown not to react with either peroxide or superoxide over shorter timescales. It is worth noting that this may change with longer-term oxidative stress.50,51,52

Lastly, some cytosolic amino acid pools are used to drive the uptake of other amino acids, meaning that a reduction in the cytosolic pool of one amino acid may lead to a deficiency in another seemingly unrelated member of the family. Indeed, the upregulation of amino acid transporters in cancer maintains amino acid pools at a level that supports its malignant hallmarks.53 This can be highlighted by a number of examples: the export of glutamine to drive leucine uptake,54 asparagine for serine, arginine and histidine55 and glutamate for cystine.56 Cystine is the oxidised homodimer of cysteine, connected via a disulfide bridge, and is the form of the amino acid found in the extracellular space. It enters the cell via a member of the heteromeric amino acid transporter family, known as the system xC– transporter, which consists of the SLC3A2-encoded CD98 protein covalently linked to the SLC7A11-encoded xCT.57 The former is required for cell-surface expression‎, while xCT is responsible for the specific carrier activity. In order to transport cystine into the cell, the transporter antiports glutamate out of the cell, functionally linking these two amino acids, which alongside glycine are critical for the synthesis of the major antioxidant, glutathione.58,59 As cellular cysteine availability is the rate-limiting step in glutathione synthesis,59 the function of the xCT transporter, and therefore intracellular glutamate concentrations, are critical for the synthesis of this major cellular antioxidant. Accordingly, as glutathione concentrations can determine the response of tumours to therapy, xCT has been suggested as a putative therapeutic target.57,60,61 There is significant demand for glutamate within a proliferating cell not only as an amino acid per se, but also from numerous transamination reactions, use of its carbon backbone for the synthesis of other anabolic metabolites and antioxidant synthesis. Interestingly, this suggests that glutamate could become limiting for proliferation in some conditions. It has been suggested that glutamate could be synthesised through the action of glutamate dehydrogenase on cellular α-ketoglutarate, recycling the ammonia produced by glutaminase activity and reducing NAD+ in the process (Fig. 1).62 Where glutamate is limiting, this reaction may therefore support tumour growth – a pathway shown to be relevant in breast cancer.62

Glycine, the third amino acid required for glutathione synthesis, is synthesised alongside serine from the glycolytic intermediate, 3-phosphoglycerate (3PG, Fig. 2) and can also be regulated by a coupled metabolic reaction. While the synthesis of serine requires both the generation of cytosolic NADH and glutamate (Fig. 2), the synthesis of glycine from serine is used to provide methyl groups for both nucleoside synthesis and regeneration of S-adenosylmethionine (SAM), required for methylation of proteins and DNA.63 It was also recently suggested that serine and glycine metabolism within cancer cells includes the use of a mitochondrial pathway, which, through the catabolism of serine to formate (which is excreted), generates mitochondrial NADH methyl groups for nucleotide synthesis.64 This pathway was since shown to depend on mitochondrial NADH oxidation (and therefore ATP synthesis).65 In this way, it represents another means of translocating cytosolic NADH into the mitochondrion, similar to the malate–aspartate shuttle, while also producing one-carbon intermediates for cellular anabolism.

In summary, it is apparent that under certain conditions, the synthesis of amino acids can be used by cancer cells to drive the other products of these reactions – whether to preserve redox homoeostasis, maintain α-ketoglutarate concentrations or for the import of other amino acids.

Non-proteinogenic amino acids in cancer metabolism

Those amino acids that can be used for protein synthesis are in fact a small subset of the total amino acid pool in the mammalian metabolome. Alongside those that are dimers of proteinogenic amino acids (including cystine and cystathionine), there are those with a β, γ or δ rather than α chiral centres (such as taurine, β-alanine and δ-aminolevulinic acid) and those with a proteinogenic structure that are not utilised in de novo protein synthesis (e.g. N-acetyl-l-glutamate, ornithine and citrulline).

Taurine is either synthesised from methionine (via cysteine) or taken up in the diet. It has several suggested functions, including efficient loading of amino acids onto some mitochondrial tRNAs in free radical scavenging and regulation of cellular osmolarity.66,67,68 It is transported across the plasma membrane by SLC6A6 (TauT), which also transports β-alanine, and by SLC36A1 (PAT1), which transports amino acids such as alanine, glycine and serine, as well as other substrates.69,70 There have been reports that taurine levels in vivo are increased in specific tumour types in the brain,71,72 and that it may play a role in the induction of cancer cell death.73 However, the main role of taurine in either health or disease is as yet unknown.

The metabolism of δ-aminolevulinic acid (5-aminolevulinic acid, 5ALA) is also clearly altered in some tumour types but has a yet-undefined role. 5ALA is the substrate for porphyrin synthesis, which forms the basis of successful responses to photodynamic therapy used to treat some tumour types, and is increasingly used to guide surgery – particularly in patients with glioma.74 The transporter for this amino acid is as yet unclear. SLC6A6, SLC6A13 (GAT2), SLC15A1 and SLC15A2 (PEPT1 and 2) have all been suggested, with the likelihood being that this function is tissue-specific.75,76,77 The fact that some tumour types take up and metabolise this amino acid at significantly greater quantities than the surrounding tissue – the basis of its use as a clinical imaging agent – suggests that we understand little of tumour requirements and use for this amino acid.

With these examples, it is clear that many of the non-proteinogenic amino acids have been highly understudied thus far, despite the fact that indications in patients suggest that they may play a role in the biological underpinning of tumours. The lack of investigations is most likely due to their absence in tissue culture medium, despite being present at often significant concentrations in tissues, plasma and cerebrospinal fluid. Indeed, both the recently described plasma-like culture media contain many of these metabolites,78,79 indicative of an increasing awareness that our common culture systems do not adequately address many aspects of cancer cell biology, forcing the metabolic network into an aberrantly small space and reducing our ability to efficiently translate novel agents to target cancer metabolism into patients.

Summary

Investigations thus far into the metabolism of amino acids in cancer has begun to highlight a nuanced network, where in many cases, tracing of the uptake and use of a specific amino acid may not reveal the real functional output of the pathway. Tumours silence expression‎ of genes required for synthesis of some amino acids, producing an auxotrophy that requires metabolic symbiosis with the stroma. However, the stroma is also highly relied upon to provide additional nutrients to support viability and proliferation, with the cancer cells only sometimes providing something to the stromal cells in return. This clearly describes a metabolic network that requires modelling approaches that take into account multiple cell types as well as physiologically relevant media. Indeed, as these media contain ‘plasma-like’ concentrations of proteinogenic amino acids, the direction and magnitude of fluxes observed are more likely to represent those observed in patients, which can therefore be more reproducibly translated into efficacious agents that can eventually improve patient care.

References

1.  
2.  
3.  
4.  
5.  
6.  
7.  
8.  
9.  
10.  
11.  
12.  
13.  
14.