Targeted metabolomic profiles of serum amino acids and acylcarnitines related to gastric cancer

Introduction

Gastric cancer (GC) represents a malignancy originating from the epithelium of the gastricmucosa (Lai, 2019). It ranks fifth and third among cancers in terms of incidence and mortality, respectively (Bray et al., 2018). In China, an estimated 679,100 incident GC cases are detected yearly, and GC represents the second deadliest malignancy in both men and women (Chen et al., 2016). Gansu province, located in the Northwest of China, represents an area with high GC incidence and mortality (Guo et al., 2020). As such, further studies of GC in this area are warranted. The majority of early GC cases show no overt symptoms. In some cases, nausea, vomiting, and/or other upper gastrointestinal tract symptoms are observed, which are not specific (Gong & Zhang, 2020). Most GC cases are diagnosed at an advanced stage, owing to the disease being mostly asymptomatic in the early stage as well as the unavailability of adequate screening methods. At present, GC treatment mainly relies on surgery, in combination with radiation therapy and chemotherapy (He et al., 2020). Surgery followed by chemotherapy in GC increases survival in the early stage, but not in the intermediate and advanced stages (Zhou et al., 2020).

Early detection and timely treatment following preciserisk classification are imperative for improving patients outcome in GC. Endoscopic assessment represents a very reliable method, which could perform GC diagnosis but is expensive and carries inherent risks related to all invasive procedures, limiting its clinical applicability (Chan et al., 2014). Conventional serum tumor biomarkers have been proposed as useful parameters for early detection, prognostic prediction and recurrence monitoring in GC (Emoto et al., 2012). Nonetheless, most circulating molecular markers are not recommended for early GC diagnosis, with limited specificity and sensitivity (Wu et al., 2019). Cell-free nucleic acids (cfNAs) are potential biomarkers for early detection and monitoring of GC (Wan & Zhang, 2016). However, their heterogeneous origin differences in specimen collection and storage, diverse assessment techniques; and various analytical methods may all affect the levels of cfNAs to limit their use as biomarkers (He et al., 2015). The identification and validation of GC-specific biomarkers are therefore warranted to facilitate early diagnosis.

GC has a multifactorial etiology, which is influenced by several genetic, environmental and predisposing factors (Agkoc et al., 2010). It is characterized by a stepwise progression from no active gastritis to chronic active gastritis, precursor GC lesions (atrophy, intestinal metaplasia and dysplasia), and finally gastric adenocarcinoma (Kuligowski et al., 2016). Atrophic gastritis (AG) is a precancerou scondition, with elevated risk of developing gastricneoplasias, including intestinal adenocarcinomas and gastric neuroendocrine tumors (Annibale, Esposito & Lahner, 2020). Therefore, it is necessary to observe the progression from gastritis to gastric cancer, and identifying diagnosis biomarkers for GC is of great significance.

Metabolomic analysis is a promising approach that provides opportunities to elucidate complex tumor-associated metabolic changes and accelerates the process of identifying new tumor biomarkers (Aboud & Weiss, 2013). Metabolic ailments are critical in carcinogenesis and cancer progression, and considered important hallmarks of cancer (Lario et al., 2017). Metabolomic profiling might help determine cancer markers, providing clues for early cancer diagnosis. Metabolomics is considered a rapid and efficient tool for identifying new cancer biomarkers, and serves as a complementary method to gene and proteome analyses (Corona et al., 2012). The serum is commonly considered a pool of metabolites (López-Bascón et al., 2016) and reflects the systemic metabolic regulation in cancer patients. Multiple methods have been developed for metabolome assessment, including nuclear magnetic resonance spectroscopy (NMR), mass spectroscopy (MS), and liquid (LC) and gas (GC) chromatography (Chan et al., 2014). To date, LC-tandem mass spectroscopy (LC-MS/MS)-based high throughput techniques are broadly utilized, allowing joint assessment of multiple metabolites, including amino acids and acylcarnitines, only requiring small amounts of biological specimens (Sarker et al., 2019).

Dysregulated metabolism of glucose, amino acids, lipids, and nucleotides has been demonstrated during gastric carcinogenesis (Huang et al., 2020). It is currently admitted amino acids, in lieu of glucose, are the main drivers of carbon-based biomass production in fast-growing malignant cells. In addition, amino acids have nitrogen, representing the major nitrogen source for hexosamines, nucleotides and other nitrogenous molecules in fast-growing cells (Choi & Coloff, 2019). Acylcarnitines constitute obligate cofactors in mitochondrial fatty acid β-oxidation, which represents the major step of energy production and is altered in multiple malignancies (Lu et al., 2019). Excessive acylcarnitine production might reflect altered fatty acid oxidation, in turn contributing to metabolic diseases. In cancer, acylcarnitine metabolism is considered a gridlock to precisely promote metabolic flexibility based on its crucial function in controlling the metabolic processes of glucose and fatty acids. Metabolic reprogramming in malignant cells modulates the production of acylcarnitines of different chain lengths. This intercalates acylcarnitines with other major metabolic pathways, factors and metabolites, resulting in balanced energy production and consumption, as well as in the biosynthesis of metabolic intermediates for fast growth (Li, Gao & Jiang, 2019). Amino acids and acylcarnitines are potential biomarkers for cancer diagnosis, with critical roles in cell physiology as baseline metabolites and metabolic modulators. However, studies evaluating amino acid and acylcarnitine profiles in association with gastric cancer are scarce and inconsistent. Jing et al. (2018) compared 22 plasma amino acids of gastric ulcer and gastric cancer patients, and demonstrated four amino acids (glutamine, histidine, arginine and tryptophan) show reduced amounts, whereas ornithine content was increased in plasma specimens from patients with GC. Miyagi et al. (2011) observed the difference of 19 plasma amino acids levels between normal people and gastric cancer patients and found that threonine, glutamine, alanine, citrulline, leucine, phenylalanine, lysine, asparagine, methionine, citrulline, valine, tryptophan, histidine and arginine were decreased in cancer. However, another study found that plasma arginine level in the gastric cancer group was significantly higher than that in the gastric ulcer, gastric polyp, and gastritis groups (Shi et al., 2021). A similar situation was found for acylcarnitines, Corona et al. (2018) found that serum acetylcarnitine (C2), hexadecanoylcarnitine (C16) and octadecenoylcarnitine (C18:1) of GC patients were higher than first-degree relatives. A study performed by Lario et al. (2017) found that plasma hydroxytetradecadienylcarnitine (C14:2-OH) and octadecanoylcarnitine (C18) were increased in GC patients compared to non-active gastritis (NAG), chronic active gastritis (CAG) and precursor lesions of gastric cancer (PLGC). Therefore, more studies are required to assess metabolic changes in gastric cancer.

The present work primarily aimed to observe changes in serum metabolomics during the whole process from superficial gastritis (SG) to AG eventually progressing to GC, and to identify a potential biomarker panel to distinguish GC from SG and AG.

Materials & Methods

Results

Discussion

Gastric cancer (GC) is among the most common cancers worldwide, which is considered a multistage progressive process (Kuligowski et al., 2016). AG is currently assessed as a pre-malignant gastric lesion in patients at risk of progression to GC, as expected in the Correa’s cascade model, which is a widely accepted GC pathogenesis model (Annibale, Esposito & Lahner, 2020). New noninvasive biomarkers of GC has become an active field of research that would improve the detection rate of early GC and therefore the prognosis of this prevalent neoplasia.

In this study, we investigated the 42 metabolites variation using serum from patients with GC, AG and SG. Our experimental found changes in serum metabolomics during the whole process from SG to AG and then to GC, and constructed a panel of metabolomic markers including C6DC, C16OH, C6, C0 and ARG that was suitable for the diagnosis of GC. Moreover, our results showed that the diagnosis model sensitivity and specificity in GC diagnosis were 98.55% and 99.32%.

As shown above, GC cases had elevated plasma ARG levels compared with SG and AG patients, in agreement with a study examining gastric juices amples and a study examining plasma ARG (Liu et al., 2018; Shi et al., 2021), but in contrast with Jing et al. (2018). ARG contributes to the metabolism of nitric oxide (NO), a vasodilator and free radical involved in inflammatory reactions and cancer development via nitro-oxidative stress, apoptosis, cell cycle, angiogenesis, invasion and metastasis (Ni et al., 2019). High ARG amounts were considered to be responsible for NO elevation (Yang, Taboada & Liao, 2009) Inducible nitric oxide synthase (iNOS) was suggested to promote NO production in GC initiation and progression (Yang, Taboada & Liao, 2009). Moreover, high levels of argininosuccinate synthase 1 (ASS1) induce tumor growth and aggressiveness by elevating ARG amounts for NO synthesis (Keshet & Erez, 2018). High NO flux leads to the formation of reactive nitrogen species (RNS) that causes deoxyribonucleic acid (DNA) damage and/or mutations, which could eventually induce carcinogenesis in gastric cancer (Wink et al., 1991). Also, the NO flux can induce cMyc mutation through the EGFR pathway (Somasundaram et al., 2019) (Fig. S2). Results of this study demonstrated that alanine, citrulline, leucine, methionine, phenylalanine, proline, tyrosine, valine, glycine and ornithine could not distinguish GC from AG and SG. This result differed slightly from those of previous studies. Miyagi et al. (2011) reported threonine, glutamine, alanine, citrulline, leucine, phenylalanine, lysine, asparagine, methionine, citrulline, valine, tryptophan, histidine and ARG amounts are reduced in GC patients, ornithine, proline, glycine levels were not notably different between GC and healthy group (Miyagi et al., 2011). Jing et al. (2018) studied the amino acid profiles of gastric cancer and gastric ulcers patients, and found glutamine, histidine, ARG and tryptophan reduced amounts, whereas ornithine content was increased in plasma specimens from patients with GC. They furtherly established regression model showed AUC at 0.922 with 85.5% specificity and 89.1% sensitivity. The most likely reason for differences between our findings and previously published studies is the different criteria for sample grouping.

Acylcarnitines enter the mitochondria, where they are converted by CPT2 into their CoA esters, which then undergo β-oxidation (Houten & Wanders, 2010). As such event is important in fatty acid metabolism and intracellular CoA homeostasis, acylcarnitine accumulation is associated with insufficient fatty acid oxidation and reduced mitochondrial function (Adeva-Andany et al., 2019) (Fig. S3). Intracellular acylcarnitine amounts might reflect serum levels (McGill et al., 2014). Here, we found that C6DC, C16OH, C6 and C0 amounts were elevated in GC cases. Upregulation of the β-oxidation process might increase C6DC, C16OH, C6 and C0 amounts. Therefore, reducing in C6DC, C16OH, C6 and C0 concentrations maybe a novel approach for cancer therapy. Corona et al. (2018) reported GC patients are characterized by increased amounts of acylcarnitine derivatives (C2, C16 and C18:1), corroborating the current findings.

Furthermore, ROC curve was built in the present study to evaluate the diagnostic values of the semetabolites. The combined detection factor had a higher area under the ROC curve compared with each single metabolite. The AUC of the combined detection factor was 0.9977, with specificity and sensitivity of 98.55% and 99.32%, respectively. This indicated that the diagnostic performance of the combined metabolites was obviously higher than that of each single metabolite. Moreover, the combined diagnostic panel had a favorable diagnostic effect, providing a new basis for the early diagnosis of GC. However, this was a single-center trial including few patients. Future trials with larger patient cohorts are needed to substantiate the above results.

Conclusion

In summary, increased C6DC, C16OH, C6, C0 and ARG expression distinguished GC from AG and SG, indicating C6DC, C16OH, C6, C0 and ARG are potential diagnostic markers of GC. This study was based on observed metabolite changes during the whole process from SG to AG and then to GC. Thus, our study provides insights in understanding metabolite profiles of amino acids and acylcarnitines also associated with gastric cancer and their use as cancer biomarkers. Further studies with larger patient cohorts to further assess clinical relevance are needed as patients is still limited.

Supplemental Information