Impact of Age on Tumor Mutation Burden in Gastric patients:a secondary analysis based on a cross-sectional study

Background Gastric carcinoma is aggressive cancer with a poor prognosis. Recent years, immunotherapy has been validated effective in a proportion of patients. Tumor mutational burden (TMB),microsatellite instability (MSI) status and programmed cell death-ligand 1 (PD-L1) expression are potential biomarkers to predict efficacy to immunocheckpoint inhibitors (ICIs). This study was undertaken to explore the correlationship between these biomarkers using the published data in a large patient cohort of gastrointestinal cancers. Method In total,367 gastric cancer patients were examined in this study. MSI and TMB were assessed by next-generation sequencing (NGS). PD-L1 expression was determined by immunohistochemistry. The population was divided into four groups according to age,TMB-High and MSI-high participants were more frequent in older groups.After adjusting sex, differentiation,histology,specimensite,MSI status and PD-L1 expression, a non-linear relationship between age and TMB was found,which had an inflection point of 73 years.The effect sizes and the confidence intervals on the left and right sides of the inflection point were 0.06 (-0.01to 0.13) and 0.56 (0.33 to 0.79), respectively. In different subgroups analysis,similar trends were observed. The relationship between age and TMB is non-linear. Age was positively correlated with TMB when age was more than 73 years.


Background
Gastric carcinoma is aggressive cancer with a poor prognosis. Recent years, immunotherapy has been validated effective in a proportion of patients. Tumor mutational burden (TMB),microsatellite instability (MSI) status and programmed cell death-ligand 1 (PD-L1) expression are potential biomarkers to predict efficacy to immunocheckpoint inhibitors (ICIs). This study was undertaken to explore the correlationship between these biomarkers using the published data in a large patient cohort of gastrointestinal cancers.

Method
In total,367 gastric cancer patients were examined in this study. MSI and TMB were assessed by nextgeneration sequencing (NGS). PD-L1 expression was determined by immunohistochemistry.

Results
The population was divided into four groups according to age,TMB-High and MSI-high participants were more frequent in older groups.After adjusting sex, differentiation,histology,specimensite,MSI status and PD-L1 expression, a non-linear relationship between age and TMB was found,which had an inflection point of 73 years.The effect sizes and the confidence intervals on the left and right sides of the inflection point were 0.06 (-0.01to 0.13) and 0.56 (0.33 to 0.79), respectively. In different subgroups analysis,similar trends were observed.

Conclusion
The relationship between age and TMB is non-linear. Age was positively correlated with TMB when age was more than 73 years.
Despite a steady decline, gastric cancer (gastric cancer) remains the third leading cause of cancer mortality in the world and a serious health burden in the world [7]. In the last few years, the immune checkpoint inhibitors (ICIs) have gained impetus in the treatment of advanced gastric cancer, and PD-L1 expression is considered as a potential predictive biomarker for efficacy to anti-PD-1 therapy [8].
However, more and more studies showed that PD-L1 expression does not fully explain the overall survival benefit gained from these ICIs. In order to maximize the clinical response for patients, more predictive and prognostic biomarkers should be explored,such as tumor mutation burden (TMB), microsatellite instability (MSI) and Epstein-Barr virus (EBV) status [9][10][11][12].
In a large patient cohort of gastrointestinal cancers,TMB,MSI and PD-L1 expression were quantified and determined their interrelationship in gastrointestinal cancers.In their paper, the authors found that,MSI is the primary driver for TMB-high, and TMB-high rate varied widely among gastrointestinal carcinomas [13]. In this paper, we performed a secondary data analysis based on aforementioned data using the subset 'gastric cancer' patients. Measurement of TMB,MSI and PD-L1 expression TMB, as measured by whole exon sequencing and counting all nonsynonymous missense mutations per tumor sample which haven't not been previously described as germline alterations,is associated with clinical benefit from immunotherapy. An Agilent SureSelect XT assay (Agilent, Santa Clara, CA,592 genes and 1.4 MB sequenced/tumor) was used to calculate TMB. The threshold to define TMBhigh was greater than or equal to 17 mutations/MB and was established by comparing TMB with MSI by fragment analysis in colorectal cancer cases.MSI was detected using over 7,000 target microsatellite loci and comparing with the reference genome hg19 from the University of California, Santa Cruz (Santa Cruz, CA) Genome Browser database. The number of microsatellite loci resulting from somatic insertion or deletion was counted for each tumor sample. Only insertions or deletions which would increase or decrease the number of repeats were considered to be of significance.

Methods
Genomic variants in the microsatellite loci were detected using the same method in mutation detection. MSI results were compared with results from over 2,000 matching clinical cases analyzed with traditional PCR-based methods. The threshold to determine MSI was set to be≥46 loci with insertions or deletions to generate a sensitivity of >95% and specificity of >99%. Automated staining techniques with SP142 (Spring Biosciences) as primary antibody was used to assess PD-L1 expression.The staining intensity on the membrane of the tumor cell was scored according to a semiquantitative scale of 0-3( 0 for no staining, 1 for weak staining, 2 for moderate staining, and 3 for strong staining) and the percentage of positively stained cells was 5% [13].

Statistical analysis
Continuous variables were expressed as median (inter-quartile range), and categorical variables were expressed in frequency or as a percentage. Kruscal Whallis H test and chi-square tests were applied t o detect the statistical differences between different groups. Generalized additive model (GAM) was u sed to identify the non-linear relationship between age and TMB. If the non-linear correlation was foun d, a two-piecewise linear regression model was used to examine the threshold effect of the age on TM B in terms of the smoothing plot. The subgroup analyses were performed using the calculated thresho ld, and interaction between subgroup was also inspected. All of the analyses were performed with the statistical software packages R (http://www.R-project.org, The R Foundation) and EmpowerStats (http: //www.empowerstats.com, X&Y Solutions, Inc., Boston, MA). P values less than 0.05 (two-sided) were c onsidered to be statistically significant.

Tumor characteristics
In total, this cohort was composed of 143(38.96%) female patients and 224(61.04%) male patients.Th e median age was 62 years(IQR,51-73).Most specimens obtained from primary tumor sites (64.85%) c omparing with the metastatic sites (35.15%). 330 samples were classified to adenocarcinoma,and the remaining 37 cases included 1 mixed adenocarcinoma/large cell neuroendocrine carcinoma,4 neuroen docrine carcinoma and 6 adenosquamous carcinoma,the other 26 cases can't be divided into any gro up of the above. In the cohort,155 tumor samples were recorded as "poorly differentiated", 29 cases " moderately differentiated",24 cases "moderately to poorly differentiated",1case "well to moderately d ifferentiated",the remaining 158 cases without specific records.
About the TMB examination, as 17 mutations/MB the cutoff point, TMB-High group comprised 8.45%of the cohort. About the MSI and PD-L1 results, 29 cases were classified as MSI-High and 34 cases were classified as positive group(Table1). Noted:IQR,inter-quartile range.
Interrelationship between TMB, MSI, and PD-L1 was displayed in Figure 1. It was shown that the distribution of TMB-high group was highly consistent with that of MSI-high,but not with PD-L1 positive cases.Therefore, the integration of TMB or MSI would broaden the potentially beneficial population who might response to ICIs.  2).

Discussion
In recent years, cancer immunotherapy has revolutionized the oncology landscape by rebalancing the host immune system. Blocking immune checkpoints such as cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4)and PD-1/PD-L1, has proven effective in a series of solid malignance [16]. Recent studies suggested that ICIs might become a future therapeutic option in gastric cancer. Based on results from KEYNOTE-059 trial, Food Drug Administration (FDA) approved pembrolizumab after two or more prior lines of therapy for patients whose tumor overexpressed PD-L1 in advanced gastric and gastroesophageal junction cancer [17].However, the relationship between PD-L1 immunohistochemical positivity and response to ICIs have been shown almost invariably in patients with either metastatic melanoma or Non-small Cell Lung Cancer(NSCLC) [18].Because gastric cancer also belongs to epithelial malignancies with many similarities in etiology, morphology and mutational profile with NSCLC, it is rational to extrapolate the correlations between PD-L1 overexpression and ICIs response.
In KEYNOTE 059 and Checkmate-032 studies , the objective remission rate(ORR) and Disease control rate(DCR) were respectively higher in PD-L1+ tumors,however, in KEYNOTE-061 and JAVELIN GASTRIC 300 studies, no clinical improvement was observed in PD-L1 + tumors [17,19]. In all, the accessible data suggest that PD-L1 may not be the perfect biomarker for selecting patients in gastroesophageal cancer patients for immunotherapy. Therefore,more studies predicting immunotherapy efficacy are going ahead, such as MSI-status, EBV positivity , tumor infiltrating lymphocytes (TILs),TMB and so on [20][21][22].Taking any of these indicators into consideration would hypothetically maximize the clinical response of a specific subgroup of patients.
In Salem et al study, TMB, MSI-status and PD-L1 expression were examined, as in Figure 1, the distri bution of MSI-status was obviously more consistent with TMB than PD-L1 expression [13].Even though MSI-status was tightly associated with TMB, it was possible to observe TMB-high subjects in MSS tumo rs [23].Therefore, the combination of TMB,MSI-status and PD-L1 expression would broaden more poten tial subjects who might benefit from immunotherapy.
Findings in selected cancer types suggest that TMB may predict clinical response to ICI [23]. In this secondary analysis, the impact of age on TMB in gastric cancer patients was first founded.
As far as we know, there were limited studies about the impact of age on TMB. Similarly, Wang et al r eported that younger gynecologic cancer patients (age <40 years) had a significantly lower TMB than older patients (age ≥40 years) [25]. In an initial clinical cohort of 102,292 samples for 167 distinct can cer types, a significant increase in TMB associated with increased age was documented [26].This phen omenon could be explained by age-related somatic mutations in cancer genome. Studies in mice and fruit flies have shown that somatic mutations accumulate with age in an issue-specific style, with resp ect to both the rate of age-related increase and the types of mutations found to accumulate [27,28]. T he report from a 6,969 samples' study showed that the number of somatic mutations in tumors was di stinctly higher when the tumor derived from an old patient, comparing with a young one [29,30].The molecular biological mechanisms of aging involving somatic mutation may include p16(Ink4a) mutatio n,dysfunction of P53,DNMT3A mutations,SIRT1 gene disorder and some unspecified mechanisms [31-3 4]. TMB calculation is conducted by counting all nonsynonymous missense somatic mutations. So we t hought it is plausible that TMB could mirror the somatic mutation in genome. As far as we know, this i s the first report to show the influence of age on TMB in gastrointestinal cancers. It gives us some hint that, immunotherapy should be given priority to older patient when anti-tumor decision making,comp aring to young patients. But the correlation might vary to some extent between cancer types, for exa mple,Puccini et al found an opposite result that higher rates of TMB-high tumors occur in younger pati ents than in old patients [14].So there maybe some other unilluminated mechanism to conceal the age -related increase in mutation burden.
There are some limitations in this study. Firstly, due to the nature of cross-sectional study, we pro vide only weak evidence between age and TMB, and it is difficult to distinguish cause and effect. Seco ndly, the tumor sample collected is restricted to the American, so the generalizability is restricted. Las tly, this study is a secondary analysis based on published data,so variables not included in the raw da tasheet cannot be analysed, such TNM stage, EBV status and Lauren type.

Conclusion
A non-linear relationship between age and TMB was found in this manuscript,which had an inflectio n point of 73 years, after adjusting sex, differentiation,histology,specimensite, MSI-status and PD-L1ex pression.As far as we know, this is the first report demonstrating the correlation between age and TM B in smooth curve in various types of tumors. So we think that immunotherapy should be considered f or patients when anti-tumor scheme making.

Declarations
Availability of data and materials The dataset supporting the conclusions of this article is available from: Landscape of tumor mutation l oad, mismatch repair deficiency, and PD-L1 expression in a large patient cohort of gastrointestinal ca ncers.https://datadryad.org/resource/doi:10.5061/dryad.qt3v0t4).

Authors'contributions
The study was conceived and designed by Baolan Li,Tongmei Zhang and Mingming Hu. Mingming Hu,analyzed the data and drafted the manuscript. Jie Li and Yuan Yang preliminary reviewed the manuscript. All authors have read and approved the manuscript.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests. The relationship between age and TMB. A nonlinear relationship between them was detected after adjusting for sex, differentiation,histology,specimensite, MSI -status and PD-L1 expression.