Remember me
ORIGINAL ARTICLE
Year : 2023 | Volume
: 41
| Issue : 2 | Page : 87-93
Expression of acyl-CoA synthetase medium-chain 3 is associated with obesity in melanoma patients and correlates with androgen receptor
Yuan Zheng, Pingdong Jiang, Liyin Zhang
Department of Dermatology, Wuxi No. 2 People's Hospital, Wuxi, China
Correspondence Address:
Dr. Yuan Zheng
Department of Dermatology, Wuxi No. 2 People's Hospital, 68 Zhongshan Road, Wuxi 214002, Jiangsu
China
Source of Support: None, Conflict of Interest: None
DOI: 10.4103/ds.DS-D-22-00141
Background: Malignant melanoma is a typical type of cancer that feature the obesity paradox. Objectives: We aim to evaluate the role of acyl-CoA synthetase medium-chain 3 (ACSM3) in obesity paradox in malignant melanoma (MM). Methods: With reproduction of the Cancer Genome Atlas (TCGA) MM dataset and validation using immunohistochemistry, we evaluated correlations of ACSM3 expression with body mass index (BMI), gender, and androgen receptor (AR) expression. Results: ACSM family genes were downregulated in MM and in normal skin exposed to ultraviolet. Higher expression of ACSM3 was associated with improved overall survival in men but not in women. Subgroup analysis showed the best survival outcome in obese patients with high ACSM3 expression. ACSM3 expression was significantly lower in cases with ulceration. NET-GE-based gene enrichment analysis of ACSM-overexpressed cases showed significant enrichment in lipid metabolism and butanoate metabolism. AR stood as the top possible transcription factor of ACSM3 using multiple algorithms. Expressions of ACSM3 and AR were positively correlated in obese men and overweight women. ACSM3 expression was positively correlated with BMI in men and overweight women. Conclusion: ACSM3 expression is associated with obesity in MM patients and correlates with AR. Functional analysis linking the findings to the obesity paradox warrants further study.
Keywords: Acyl-CoA synthetase medium-chain 3, malignant melanoma, obesity paradox
Malignant melanoma (MM) is highly aggressive taking up of 90% of deaths caused by cutaneous malignancy.[1] Unlike most solid cancer, MM can occur at or metastasize to almost any site of the body. Globally, cutaneous MM consists of 1.7% of newly diagnosed malignancy and accounts for 0.7% of cancer-related mortality, with an increased trend in incidence.[2] Asia has a lower incidence for MM, but the 5-year survival in China is only 65%, lower than ~80% in Western countries.[3] In general, MM confers a poor prognosis despite emerging novel treatments.[4]
Thus far, there has been no curative treatment for advanced MM. The mainstay of treatment approved by the FDA includes MAPK-targeted therapy (BRAFi, MEKi, etc.) and immune checkpoint inhibitors (ICI).[5],[6],[7],[8],[9] Over 50% of metastatic MM cases harbor BRAFV600E mutation. In the COMBI-AD trial that recruited stage III MM patients with BRAFV600E/K who underwent radical excision and lymph node dissection, a combination of BRAFi and MEKi conferred a 53% reduction in recurrence and death as compared to the control. However, the intriguing phenomenon was observed alongside the promising effectiveness of targeted therapy.
The obesity paradox refers to the phenomenon that obesity is associated with a higher incidence of certain cancer yet is meanwhile associated with improved clinical outcome. MM and renal cell carcinoma (RCC) are two typical types of cancer that feature the obesity paradox, being validated in a variety of studies.[10],[11] Of note, the paradox in MM is not only observed in patients undergoing extirpative treatment but is also present in cases undergoing the latest targeted and immunotherapy.[12] Although the underlying mechanism of the obesity paradox remains elusive, a few studies tend to identify genetic disposition. For example, aberrant expression of FASN was reported to mediate obesity paradox in part in RCC.[13] In MM, however, there is a dearth of studies reporting candidate genes.
Gender also plays a role in the prognosis of MM. It is established that female MM patients perform significantly better than men do, with underlying genetic dispositions also elusive. To certain extent, the obesity paradox even overtakes gender disparity showing very limited prognostic benefit in obese women than in men. Nonetheless, the survival in those obese patients almost doubled compared to their counterparts with lower body mass index (BMI). The confounding interplay between gender and obesity paradox reflects the inherent distinction in fat metabolism in men and women.
In the current study, we comprehensively evaluated expressions of genes in the butyric metabolism, one of the most important pathways in fat metabolism and identified acyl-CoA synthetase medium-chain 3 (ACSM3) as the candidate gene that could link obesity paradox and gender disparity in treatment outcome in MM.
Materials and MethodsSearching algorithm
Given that the metabolism of fat and sex hormones are extensively intertwined, we postulate that an “ideal” candidate gene should simultaneously meet the following criteria: (a) The gene is differentially expressed between normal skin and MM with lower expression in tumor tissue; (b) gain of function of the gene promotes fat accumulation and is associated with improved MM survival, or vice versa; (c) expression of the gene shows the strongest correlation in obese MM patients; and (d) the survival impact diminishes in patients with lower BMI.
In silico analysis
The Cancer Genome Atlas (TCGA) MM dataset (SKCM) was reproduced. Expression of ACSM3 and clinicopathological parameters were both retrieved using the cBioPortal platform. The firehose legacy dataset of SKCM was selected, and RNA-seq samples were opted. The OncoPrint tab was used to generate expressions of ACSM family genes in paired MM tumor. An z-score of over 2 in each pair (MM vs. paired normal) was defined as overexpression. The Human Protein Atlas (HPA) dataset was reproduced to analyze ACSM family gene expression in normal skin tissue. By querying genes of interest in the dataset, we retrieved ACSM3 expression in the ultraviolet (UV)-exposed and control skin tissue. The HPA dataset was also used to compare and plot the overall survival of MM patients with higher and lower expression of ACSM family genes grouped at automatically designated cutoff values. The NET-GE platform was used to analyze the functional annotation of enriched genes in MM cases with overexpressed ACSM family genes in the TCGA cohort. The ChEA3 platform was used to analyze genes enriched in ACSM family gene-overexpressed cases in search of the distribution of transcriptional factor (TF).[14]
Sample collection and immunohistochemistry
Surgically removed MM at Wuxi No. 2 People's Hospital from 2006 to 2020 was collected. All sections were paraffinized and de-identified. Clinicopathological data were collected. A total of 84 MM sections were collected and a standard immunohistochemistry (IHC) protocol was followed. Briefly, samples were sliced at 4 μm and mounted on polylysine-coated glass slides. Endogenous peroxidase of deparaffinized sections was blocked with 3% hydrogen peroxide. After washing with PBS, all samples were fixed with citrate buffer and sealed at room temperature with 5% nonfat milk to block nonspecific binding. ACSM3 Antibody (PA5-112973) (ThermoFisher; 1:200 dilution) and androgen receptor (AR) antibody (Cell Signaling #5153; 1:500 dilution) were added and sections were incubated at 4°C overnight. Following washes with PBS, the corresponding secondary antibody was added and sections were incubated at 37°C for 15 min. HRP-conjugated streptavidin was then added and DAB (diaminobenzidine tetrahydrochloride) was added. All samples were counterstained with hematoxylin. Positive control was referenced from the HPA dataset. Samples were assessed by two experienced pathologists using the double-blind method by randomly selecting 5 high-power (×200) fields. IHC scores of each slide were analyzed semiquantitatively by the product of intensity and extensity (the proportion of the cells stained under a microscopic field). The extensity was graded as follows: 0 for 0%–10% of tumor cells stained; 1 for 11%–25% of cells stained; 2 for 26%–50% of cells stained; and 3 for >50% of cells stained. The intensity of staining was graded as follows: (1) For light yellow; (2) for dark yellow; and (3) for brown. The final value of each slide was obtained from the sum of these two factors as follows: 0 for negative (1-2), 1 for mild (3), 2 for moderate (4), and 3 for strong (5-6). For both ACSM3 and AR, staining intensity was graded from 0 (no staining) to 3 (strong staining). A staining index was calculated as the product of staining intensity and the extensity (score 0–9). The staining indices were clustered as follows: 0 = 0, 1 = 1–3, 2 = 4–6, and 3 = 9.
This study was approved by the institutional board of Wuxi No.2 People's Hospital (No. 2020-079, 2020). Informed consent was obtained from all patients, and all procedures conformed to the ethics waiver regulation by the local institutional review board.
Statistical analysis
The Graphpad v. 9 (GraphPad Software, San Diego, CA, USA) and SPSS software were used for statistical analysis IBM SPSS Statistics for Windows, Version 27.0. (Armonk, NY: IBM Corp). The Student's t-test and Mann–Whitney test were used for comparisons between the two groups as appropriate. The two-way ANOVA was used to compare the mean or median between multiple groups with subgroup comparisons. Survival was plotted with the Kaplan–Meier curve and compared using the Log–rank test. Pearson's and Spearman's correlation were used to analyze the correlation between expressions of factors and clinicopathological parameters. The Cox regression model was used to evaluate prognostic factors. P < 0.05 was accepted as statistical significance.
ResultsDown-regulation of acyl-CoA synthetase medium-chain family genes is related to malignant transformation
We first queried the expression of ACSM3 genes in MM patients and found that overexpression was the sole alteration indicating the gain of function of ACSM family genes in MM [Figure 1]a. Of note, when we plotted expressions of ACSM family genes, we found that expressions were significantly lower in MM, and even UV exposure in normal skin could significantly downregulate ACSM1 and ACSM3 [Figure 1]b. In survival analysis, only higher expression of ACSM3 showed significantly improved overall survival [Figure 1]c. Subgroup analysis showed men with higher ACSM3 expression had improved survival compared with men with low ACSM3 expression [Figure 1]d. As expected, the best survival was observed in the cohort of obese patients with high ACSM3 expression when stratified with BMI category [Figure 1]d. We then analyzed correlations of ACSM3 expression with clinicopathological parameters and found that ACSM3 expression was significantly lower in cases with ulceration, both in TCGA and in our validation cohort [Table 1]. Of note, ACSM3 was an independent prognostic factor in MM after being adjusted for critical clinical parameters [Table 2]. ACSM3 expression was not statistically associated with other parameters, including age and mitotic count [Table 1] and [Table 3]. NET-GE-based gene enrichment analysis of ACSM-overexpressed cases showed significant enrichment in lipid metabolism and butanoate metabolism [Figure 1]e.
Figure 1: Downregulation of ACSM family genes is related to malignant transformation. (a) reproduced from TCGA MM (SKCM) dataset, shown was expression status of ACSM family genes in the MM cohort with red bars indicating cases with overexpressed ACSM gene in paired samples; (b) reproduced from Human Protein Atlas dataset, shown were expressions of ACSM family genes in healthy skin with or without UV exposure and in MM, with the lower panel showing representative IHC staining of ACSM genes in normal skin and MM; reproduced from TCGA dataset, shown were (c) Kaplan–Meier curves for overall survival in MM patients from TCGA dataset showing survival impact of ACSM family genes (P value not shown indicating > 0.05 in Log-rank test), and (d) subgroups stratified with different gender, obesity and ACSM3 expression level; (e) NET-GE enrichment analysis of genes enriched in ACSM family gene-overexpressed cases (volcano plot) showing functional annotations (*P < 0.05; ****P < 0.0001; NS = not significant; scale bar = 200 μm). ACSM: Acyl-CoA synthetase medium-chain, TCGA: The Cancer Genome Atlas, UV: Ultraviolet, MM: Malignant melanoma, IHC: Immunohistochemistry.
Table 1: Expression of acyl-CoA synthetase medium-chain 3 grouped by different clinicopathological parameters in malignant melanoma
Table 2: Multivariate analysis of the prognostic role of acyl-CoA synthetase medium-chain 3 in malignant melanoma
Table 3: Correlation between expression of acyl-CoA synthetase medium-chain 3 with body mass indexExpression of acyl-CoA synthetase medium-chain 3 is correlated with that of androgen receptor in malignant melanoma
We then tried to identify a possible transcription factor (TF) that regulated ACSM3 expression. Two datasets pointed to AR as the top hit of TFs [Figure 2]a. We further investigated the tropism of transcription activity using the CH3 platform and found liver was the most common site where the transcription occurred, corresponding to our postulation that the liver was the major site for fat metabolism [Figure 2]b and [Figure 2]c. We then studied the correlation between ACSM3 expression and BMI and found that whereas ACSM3 expression was not associated with BMI in the TCGA cohort, the expressions showed a moderate and significant linear correlation in our validation cohort [Table 3]. In men, the strongest and most significant linear correlation was observed in obese patients in the TCGA cohort and was validated in our cohort [Table 3], indicating obese men played a major role in establishing the correlation. In women, though a significant linearity was observed in general, the correlation effect was solely mild–moderate in both cohorts [Table 3]. In TCGA cohort, a significant linear correlation was only observed in obese women [Table 3]. In our validation cohort, however, a strong and linear correlation between ACSM3 and BMI was observed in overweight patients as there were not enough patients with categorized as obese, corresponding to a demographic feature of Chinese ethnicity.[15] Correlation between expressions of ACSM3 and AR was shown in obese women in our cohort and in obese men in the TCGA cohort [Table 3]. Correlation between AR and ACSM3 expression in men, in general, was unanimous in both our and TCGA cohorts [Table 3] and [Figure 2]d, [Figure 2]e, [Figure 2]f.
Figure 2: Expression of ACSM3 is correlated with AR expression in MM. (a) Queried at the Harmonizome platform, shown was transcription factors (TFs) that transcribed expression of ACSM3 in two major TF prediction datasets; (b) tissue distribution and (c) cancer type distribution of TFs regulating enriched genes in ACSM family gene-overexpressed cases profiled with ChEA3 platform; Correlations between ACSM3 expression and (d) BMI and (e) AR expression in TCGA cohort; (f) representative IHC images of expression of ACSM3 and AR in the same MM sample (scale bar = 200 μm). ACSM: Acyl-CoA synthetase medium-chain, TCGA: The Cancer Genome Atlas, MM: Malignant melanoma, BMI: Body mass index, AR: Androgen receptor, IHC: Immunohistochemistry. Discussion The obesity paradox was first proposed in the 1990s in patients with internal disorders. Obese or overweight patients who underwent dialysis showed improved survival over normal-weight patients. Later, the paradox was also documented in patients with cardiovascular diseases and was termed as obesity paradox in heart failure. In the recent decade that cancer treatment progressed rapidly, the obesity paradox has once again been observed. Besides the landmark study by McQuade et al., Naik et al.[16] also reported significantly lower mortality in male MM patients with type I obesity who received anti-PD-1 monotherapy or anti-PD-1/CTLA-4 combination therapy, similar to the outcome from an RCC study.[17]
However, most studies on obesity paradox in MM focused on clinical correlations and aimed less at functional analysis, unlike in RCC. Antonopoulos and Tousoulis reported that central and peripheral obesity that play different roles in cardiovascular diseases could be extrapolated in the obesity paradox.[18] Santoni et al. reported increased transforming growth factor-β, vascular endothelial growth factor (VEGF), and VEGF2, and decreased interferon-γ and IFN-α in obese patients with RCC.[19] Furthermore, excessive Leptin in overfed animal models of RCC could lead to increased expression of nonfunctional PD-1 on CD8+ T cells that increased response to ICI.[20] Albiges et al. have nailed down the fat metabolic gene FASN in RCC whose expression showed a negative correlation with BMI that positively correlated with prolonged survival.[13]
We have therefore extrapolated RCC studies and developed our own algorithm in search of the potential genetic disposition to obesity in MM.[21],[22] Our findings in the current study together with the report by Zhu et al., indicate ACSM3 conforms to the following features, including being differentially expressed in normal and cancerous tissue, meantime being prognostic in cancer, with opposite trending; being closely associated with obesity phenotype and fat metabolism; and highly implicated in the targeted therapy sensitivity.[23] They showed that ACSM3 expression was gradually decreased in normal, benign nevi, and MM, respectively. In MM cells, however, overexpression of ACSM3 was associated with decreased cell proliferation, halted cell cycle progression, increased cell apoptosis, decreased cell migration and anchorage-independent growth, and vice versa. Of note, they found overexpression of ACSM3 showed an additive effect in MM with BRAFi in the cell model and in vivo xenograft model. Together, such findings have been the pioneer study in the mechanistic analysis of obesity paradox in MM. In support of their findings, a more recent study by Zhao et al. further showed that KLF10 upregulates ACSM3 via the PI3K/Akt signaling pathway to inhibit the malignant progression of melanoma.[24] In the current study, we have proposed another potential regulator of ACSM3, furnishing the regulatory network of ACSM3 signaling in MM.
ACSM3 participates in the metabolism of an important exogenous short-chain fat acid, butyric acid (BA), which is produced by bacteria in the intestine. ACSM3 is thus expressed in abundance in gut and plays a role in colorectal cancer to metabolize high levels of BA, which works as an HDAC inhibitor that is detrimental to cancer cells. BA has also been recently recognized as an important factor mediating the balance of fat metabolism and is closely associated with obesity. On the other hand, ACSM3 plays distinct roles in cancers at other sites. ACSM3 has been reported to inhibit AKT signaling in hepatic cell carcinoma.[25],[26] Of note, AKT has been established as the hub mediating BRAFi sensitivity. Our findings further conform to the notion that ACSM3 plays a role in the obesity paradox in MM. We thus postulate that normally ACSM3 is downregulated in MM as it exerts anti-tumorigenic effect. In obese patients, however, the increased level of systemic BA jeopardized HDAC activity in MM, which upregulates ACSM3 in combat with BA, and the upregulation coincidentally sensitizes MM to BRAFi via AKT signaling. Furthermore, being transcribed by AR, this effect can be further augmented in men. The proposed hypothesis is now being investigated by our group. AR has been shown to play a critical role in the treatment of MM. Vellano et al. showed that AR blockade promotes response to BRAF/MEK-targeted therapy in MM in a milestone study that may alter the paradigm of MM treatment.[27] Expression of AR and activation of AR signaling have been detected in MM in a series of studies.[28],[29] Our findings on ACSM3 have bridged AR signaling with BRAFi in MM at a novel perspective.
ConclusionTo sum up, we here report that higher expression of ACSM3 is associated with obesity in melanoma patients and correlates with AR, and such a pattern corresponds to the obesity paradox. Detailed mechanism warrants further investigation.
Data availability statement
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
Comments (0)