Identification of Serum Biomarkers to Monitor Therapeutic Response in Intestinal-Type Gastric Cancer

There are a limited number of clinically useful serum biomarkers to predict tumor onset or treatment response in gastric cancer (GC). For this reason, we explored the serum proteome of the gp130Y757F murine model of intestinal-type gastric cancer (IGC). We identified 30 proteins with significantly elevated expression in early gp130Y757F IGC and 12 proteins that were significantly elevated in late gp130Y757F IGC compared to age- and gender-matched wild-type mice. Within these signatures, there was an overlap of 10 proteins commonly elevated in both early- and late-stage disease. These results highlight the potential to identify serum biomarkers of disease stage. Since IGC in the gp130Y757F model can be reversed following therapeutic inhibition of Interleukin (IL)-11, we explored whether the protein signatures we identified could be used to monitor tumor regression. We compared two different therapeutic modalities and found 5 proteins to be uniquely differentially expressed between control animals and animals halfway through treatment, with 10 differentially expressed at the end of treatment. Our findings highlight the potential to identify reliable biomarkers to track IGC tumor regression in response to treatment.


Introduction
Gastric cancer (GC) is the third most common cause of cancer-related deaths worldwide, with a higher incidence in the southern hemisphere [1,2].Many GC patients are diagnosed at a late stage of the disease, when treatments are limited to surgical resection and chemotherapy with palliative intent [3,4].The identification of reliable serum biomarkers would greatly aid early disease detection and inform treatment decisions.Indeed, serum-based GC biomarkers are utilized in routine clinical practice with the aim of early detection, disease monitoring, and improved treatment decisions; however, they have limitations [5,6].These serum biomarkers include carcinoembryonic antigen (CEA) [7], which is increased in the serum of gastric adenocarcinoma patients, but not during the early stages of carcinogenesis, and thus is not a useful marker of early disease [8].Additional prognostic serum biomarkers include CA19-9 [9] and CA72-4 [10], which are widely used in clinical practice; however, their sensitivity and thus efficacy for screening early GC is limited [5].Thus, there remains an urgent need for the discovery of reliable biomarkers.
The gp130 Y757F murine model of intestinal-type gastric cancer (IGC) [11] has been widely utilized to understand the pathogenesis of IGC and to identify and test new treatment opportunities [12].In this model, the membrane proximal tyrosine (Y) residue on the cytoplasmic tail of the transmembrane glycoprotein 130 (gp130) receptor has a phenylalanine knock-in substitution (Y757F), resulting in hyper-activation of the signal transducer and activator of transcription (STAT3) protein in response to Interleukin (IL)-6 family cytokines.The gp130 Y757F mice develop tumors in the antrum of the stomach that, like human IGC, have elevated IL-6 family cytokine and STAT3 activation, which are associated with poor prognosis in human IGC [13,14].In particular, IL-11 has been shown to be critical to disease onset and progression in the gp130 Y757F IGC model [15] and can be therapeutically manipulated to prevent and reverse tumor progression [15].Previous studies utilizing this model identified four serum biomarkers, namely afamin, clusterin, vitamin D-binding protein (VDBP), and haptoglobin (Hb), as accurate indicators of the presence of IGC in both the gp130 Y757F model and in patients, that were reported to be superior to 17].
We employed the gp130 Y757F model of IGC, as a surrogate of human disease, for the identification of potential new biomarkers of disease progression and treatment response.We performed an unbiased label-free quantitative proteomics analysis of serum collected from gp130 Y757F mice to determine if a panel of biomarkers could distinguish early from late IGC.Given the demonstrated therapeutic benefit of IL-11 signaling inhibition in this model [15], we also sought to determine if the serum protein signatures that correlated with tumor progression were altered following therapeutic manipulation of tumor burden.Given the emerging interest in biomarkers that can predict treatment response in routine clinical practice, our results identified several potential serum protein biomarkers that may inform clinical trial design.

gp130 Y757F Mice Have a Distinct Serum Proteome
We employed a label-free quantitative proteomics approach to determine if we could identify serum biomarkers that could be correlated with disease stage in the gp130 Y757F model.To this end, serum from terminal cardiac bleeds was collected from 2-month-old gp130 Y757F mice with early gastric disease for comparison with age-matched wild-type mice, as well as from 6-month-old gp130 Y757F mice with established disease for comparison with age-matched wild-type mice (Figure 1A).Principle component analysis (PCA) revealed a clear separation between the 2-month-old and 6-month-old wild-type mice, as well as between the wild-type and gp130 Y757F mice (Figure 1B).We also identified 18 proteins uniquely differentially expressed between the 2-month-old wild-type and gp130 Y757F mice (Figures 1C and S1), 21 uniquely differentially expressed proteins between the 6-month-old wild-type and gp130 Y757F mice (Figures 1D and S1), and only 2 uniquely differentially expressed proteins between the early-and late-stage disease in gp130 Y757F mice (Figures 1E and S1).
Pairwise comparison revealed that Hp, transgelin-2 (Tagln2), bridging integrator 2 (Bin2) and mannose-binding protein C (Mbl2) were significantly elevated in the serum of 2-month-old gp130 Y757F mice compared to wild-type mice and these differences were exclusive to the early GC disease stage (Figure 1C).In the 6-month-old gp130 Y757F mice with established disease, inter-alpha-trypsin inhibitor heavy chain H3 (Itih3) exhibited an increase in protein expression, while coagulation factor XII (F12) and transthyretin (Ttr) exhibited decreased protein expression compared to age-matched wild-type mice, which was unique to the late-disease stage (Figures 1D and 2A).Pairwise comparison revealed that Hp, transgelin-2 (Tagln2), bridging integrator 2 (Bin2) and mannose-binding protein C (Mbl2) were significantly elevated in the serum of 2-month-old gp130 Y757F mice compared to wild-type mice and these differences were exclusive to the early GC disease stage (Figure 1C).In the 6-month-old gp130 Y757F mice with established disease, inter-alpha-trypsin inhibitor heavy chain H3 (Itih3) exhibited an increase in protein expression, while coagulation factor XII (F12) and transthyretin (Ttr) exhibited decreased protein expression compared to age-matched wild-type mice, which was unique to the late-disease stage (Figures 1D and 2A).Comparison between the early and late disease in gp130 Y757F mice revealed an increase in ceruloplasmin (Cp), protein Z-dependent protease inhibitor (Serpina10), and serine protease inhibitor A3N (Serpina3n) in late disease, suggesting that these proteins are unique indicators of disease stage (Figure 1E).Importantly, these proteins did not exhibit significant protein expression differences in the wild-type mice when comparing early to late samples (Figure 1F).Within the comparison between 2-and 6-month-old wild-type mice, we observed an increase in the expression of 22 proteins in the 6-month-old wild-type mice, including Hp and several complement components (C8, C9) previously associated with ageing in mice [18,19] (Figure 1F).These results suggest that previous observations linking haptoglobin to IGC may be associated with age, rather than a direct reflection of disease burden [16,17].
In contrast to previous studies [16], we did not observe an increase in afamin, clusterin, VDBP, or A1AT in the serum of tumor-bearing gp130 Y757F mice versus wild-type in either early-or late-stage diseases (Figure 2A-D).However, we did find that Tagln2, Habp2, and Lifr were elevated in early disease in gp130 Y757F , while C1qa, Serpina3n, and Serpina10 were elevated in late disease (Figures 1E and 2D), indicating that they may re-present biomarkers of disease, further supported by the fact that these proteins did not change in the same comparison in wild-type mice (Figure 1F).Comparison between the early and late disease in gp130 Y757F mice revealed an increase in ceruloplasmin (Cp), protein Z-dependent protease inhibitor (Serpina10), and serine protease inhibitor A3N (Serpina3n) in late disease, suggesting that these proteins are unique indicators of disease stage (Figure 1E).Importantly, these proteins did not exhibit significant protein expression differences in the wild-type mice when comparing early to late samples (Figure 1F).Within the comparison between 2-and 6-month-old wild-type mice, we observed an increase in the expression of 22 proteins in the 6-month-old wildtype mice, including Hp and several complement components (C8, C9) previously associated with ageing in mice [18,19] (Figure 1F).These results suggest that previous observations linking haptoglobin to IGC may be associated with age, rather than a direct reflection of disease burden [16,17].
In contrast to previous studies [16], we did not observe an increase in afamin, clusterin, VDBP, or A1AT in the serum of tumor-bearing gp130 Y757F mice versus wild-type in either early-or late-stage diseases (Figure 2A-D).However, we did find that Tagln2,

Therapeutic Modulation of IL-11 Signaling Alters the Serum Proteome
Since the gastric tumors that form in gp130 Y757F mice are exquisitely sensitive to the inhibition of IL-11 signaling [15], we systemically treated age-matched mice with an anti-IL-11R monoclonal antibody (CSL-4E5) [20] or IL-11 Mutein-PEG [15] (a recombinant protein that is a potent IL-11 signaling antagonist; Figure 3A).We found that either modality of IL-11 signaling inhibition significantly reduced overall tumor burden (Figure 3B,C).The reduction in tumor burden observed was attributed to reduced epithelial proliferation (Figures 3D and S2).IL-11 is involved in megakaryocyte maturation and erythropoiesis [21].We observed no adverse effect on platelet number (Figure S3), similar to our previous reports for therapeutic inhibition of IL-11 signaling [15] indicating a favorable safety profile.
of IL-11 signaling inhibition significantly reduced overall tumor burden (Figure 3B,C).The reduction in tumor burden observed was attributed to reduced epithelial proliferation (Figures 3D and S2).IL-11 is involved in megakaryocyte maturation and erythropoiesis [21].We observed no adverse effect on platelet number (Figure S3), similar to our previous reports for therapeutic inhibition of IL-11 signaling [15] indicating a favorable safety profile.For each mouse, the serum that was collected prior to, during (mid-treatment protocol), and at the completion of the treatment protocol (Figure 3A) underwent label-free quantitative proteomics analysis to determine if a protein signature that correlated with treatment response could be identified.PCA plots demonstrated that the serum collected prior to treatment clustered together for both treatment groups, while the end-treatment IL-11 Mutein and anti-IL11R mAb groups also clustered together (Figure 4A).We then performed pairwise analyses to compare the differences in protein expression between For each mouse, the serum that was collected prior to, during (mid-treatment protocol), and at the completion of the treatment protocol (Figure 3A) underwent label-free quantitative proteomics analysis to determine if a protein signature that correlated with treatment response could be identified.PCA plots demonstrated that the serum collected prior to treatment clustered together for both treatment groups, while the end-treatment IL-11 Mutein and anti-IL11R mAb groups also clustered together (Figure 4A).We then performed pairwise analyses to compare the differences in protein expression between the pre-treatment (Pre) versus halfway through the treatment (Mid), and pre-treatment versus the completion of the treatment (End), for both the anti-IL-11R mAb-and IL-11 Mutein-treated groups compared with the control-treated gp130 Y757F mice.In the serum collected prior to treatment versus at the end of the experiment, there were 48 unique proteins differentially expressed in the serum of mice treated with anti-IL-11R mAb when compared to the control animals and 7 unique proteins that were differentially expressed in the IL-11 Mutein-treated mice when compared to the control animals (Figure S4).Within these proteins, 10 proteins overlapped between both treatment groups, including several immunoglobulins (Igh, Ighm, Igkc) and CD5 antigen-like protein (Cd5l), which exhibited increased expression in the treated mice and may be indicators of treatment success (Figure 4B,C).
We found that at the completion of therapeutic treatment with the anti-IL11R mAb, several proteins, including membrane palmitylated protein 1 (Mpp1), peroxiredoxin 6 (Prdx6), and erythrocyte membrane protein band 4.2 (Epb4.2),were significantly increased in the control-treated mice, and interleukin-1 receptor accessory protein (Il1rap), Cd5l, and several immunoglobulins were decreased (Figure 4B), suggesting that these may be markers of treatment success.At the completion of the IL-11 Mutein treatment, a significant increase in IL-11 protein was detectable in the IL-11 Mutein-treated mice compared with controltreated mice, corresponding with the recombinant IL-11 Mutein provided for treatment (Figure 4C).We also saw increased expression of coagulation factor XI (F11), tripeptidyl peptidase 2 (Tpp2), and complement component 1 (C1s1) in the control-treated mice, suggesting that their decrease in the gp130 Y757F mice was an indicator of treatment success (Figure 4C).In addition, there was an increase in Fn1, Cd5l, and platelet glycoprotein 1b (Gp1ba) in the IL-11 Mutein-treated group (Figure 4C).
the pre-treatment (Pre) versus halfway through the treatment (Mid), and pre-treatment versus the completion of the treatment (End), for both the anti-IL-11R mAb-and IL-11 Mutein-treated groups compared with the control-treated gp130 Y757F mice.In the serum collected prior to treatment versus at the end of the experiment, there were 48 unique proteins differentially expressed in the serum of mice treated with anti-IL-11R mAb when compared to the control animals and 7 unique proteins that were differentially expressed in the IL-11 Mutein-treated mice when compared to the control animals (Figure S4).Within these proteins, 10 proteins overlapped between both treatment groups, including several immunoglobulins (Igh, Ighm, Igkc) and CD5 antigen-like protein (Cd5l), which exhibited increased expression in the treated mice and may be indicators of treatment success (Figure 4B,C).We found that at the completion of therapeutic treatment with the anti-IL11R mAb, several proteins, including membrane palmitylated protein 1 (Mpp1), peroxiredoxin 6 (Prdx6), and erythrocyte membrane protein band 4.2 (Epb4.2),were significantly increased in the control-treated mice, and interleukin-1 receptor accessory protein (Il1rap), Cd5l, and several immunoglobulins were decreased (Figure 4B), suggesting that these may be markers of treatment success.At the completion of the IL-11 Mutein treatment, a significant increase in IL-11 protein was detectable in the IL-11 Mutein-treated mice compared with control-treated mice, corresponding with the recombinant IL-11 Mutein We found that within the serum collected prior to treatment versus halfway through the treatment protocol, there were no significant differentially expressed proteins within the mice treated with the anti-IL-11R mAb compared with control-treated mice (Figure S4).We did, however, observe a significant increase in the expression of proteins in the endtreatment anti-IL-11R mAb mice compared to the same mice pre-treatment, including Mpp1, Ephb4.1, and Ephb4.2 (Figure S5A).We also observed decreased expression of several immunoglobulins which was maintained at the end of the treatment (Figure S5A).Compared to the control-treated mice, there were five unique differentially expressed proteins in the IL-11 Mutein mid-treatment group (Figure S4), including Sepp1, Itih3, and Serpina3m, which were elevated in the control-treated mice compared with the Mutein-treated mice, suggesting that they may indicate treatment success (Figure 4D).Itih3 was previously found to be increased in late GC disease in gp130 Y757F mice compared to wild-type mice (Figure 1D).We also observed a significant increase in the expression of Itih3, Glo1, Anxa7, and Col1a1 in the pre-treatment Mutein mice compared to the same mice mid-treatment (Figure S5B).This further supports Itih3 as an early indicator of treatment response.In addition, we observed increased expression of Il-11, Cd5l, Gp1ba, and several immunoglobulins at the IL-11 Mutein treatment endpoint, compared to the same mice pre-treatment (Figure S5B).The increase in IL-11 corresponds to the IL-11 Mutein-PEG in circulation, while immunoglobulins are associated with stromal remodeling in early gastric cancer [22].Together with the observations in the anti-IL-11R antibody groups, these results further suggest that alterations in the tumor matrix are detectable in the serum proteome and reflect treatment response.

Discussion
GC is a devastating disease for which there are few treatment options [3].Animal models have greatly increased our understanding of the pathophysiology of the disease and enabled the study of potential new therapeutic opportunities [23].However, missing from the pipeline is the co-discovery of biomarkers to identify early disease or biomarkers to indicate a treatment is impacting tumor growth.Here, we have utilized an animal model of IGC to identify potential blood-based biomarkers unique to disease stage, in addition to biomarkers that may predict tumor regression, with a focus on therapeutic inhibition of IL-11 signaling.These studies are timely, as monoclonal antibodies targeting IL-11 or its receptor have entered Phase I/II clinical trials [24].

Serum Biomarkers of Advanced Disease
We identified serum biomarkers that were exclusive to early or late IGC (Figure 5).Each of the markers we identified have previously been linked to human cancer, and thus are not limited in relevance to the murine model employed.Of these, Tagln2 has previously been linked to angiogenesis, proliferation, migration, and epithelial-mesenchymal transition in GC [25,26], while Mbl2 gene variants have been linked to risk of stomach cancer [27] and increased serum levels linked to pancreatic cancer [28].In addition, Itih3 has previously been identified as a predictor of GC in mice and patients [29], while F12 is a prognostic marker for thyroid cancer [30].Previous studies have shown that Serpina3n levels correlate with inflammation in colon cancer [31], and promotes invasion and migration in breast cancer [32].

Comparison of Two Treatment Modalities That Block IL-11 Signaling
IL-11 has been linked to the onset and progression of numerous solid malignancies and has been gaining increased interest as a novel therapeutic target [33].In GC, IL-11 is associated with tumor progression, invasion, and poor survival outcomes [34][35][36][37].Thus,

Comparison of Two Treatment Modalities That Block IL-11 Signaling
IL-11 has been linked to the onset and progression of numerous solid malignancies and has been gaining increased interest as a novel therapeutic target [33].In GC, IL-11 is associated with tumor progression, invasion, and poor survival outcomes [34][35][36][37].Thus, utilization of the gp130 Y757F model of GC, in which IL-11 has a major role in GC formation [15], allowed for a unique opportunity to explore whether markers of tumor regression following treatment could be detected within the serum proteome.
To provide future directions to improve IGC patient selection and treatment outcome, we compared two different inhibitors of IL-11 signaling, namely a monoclonal antibody and a recombinant protein.While both inhibitors showed similar therapeutic efficacy, IL-11 Mutein, the recombinant protein revealed different markers of treatment response (Figure 5).Interestingly, our results revealed that changes in proteins associated with the tumor matrix were present following both treatment modalities and may thus represent an opportunity for the development of companion diagnostics to track tumor regression.

Serum Biomarkers of Tumor Regression
We found changes in numerous proteins at the end of the anti-IL-11R treatment compared to controls (Figure 5).Of these, Mpp1 is associated with the cytoskeleton and cell proliferation [38], Prdx6 is associated with migration and invasion in colon cancer [39], and Epb4.2 predicts survival outcomes in pancreatic cancer [40].It has also been shown that il1rap regulates apoptosis in the stomach in GC models [41].Ephb4.1 and Ephb4.2 are frequently elevated in tumors [42] and may be a reflection of residual disease.At the end of the IL-11 Mutein treatment, we observed changes in proteins such as Tpp2, which is implicated in cellular division and apoptosis and is significantly increased in squamous cell carcinomas [43], while C1s1 and Gp1ba have been linked to fibroblasts and matrix in other solid tumor types [44,45], suggesting matrix remodeling may occur following treatment.
We also compared serial serum samples from the same mouse.While we did not detect changes in response to the anti-IL-11R treatment until the end of the treatment, we did detect changes in response to the IL-11 Mutein treatment throughout the treatment cycle.This included changes in Glo1 and Anxa7 which have previously been shown to be elevated in GC tumors [46].Col1a1 is overexpressed in GC [47], and its reduction following IL-11 Mutein treatment further suggests remodeling of the tumor matrix.

Strengths and Limitations
By comparing healthy animals to treatment-naïve gp130 Y757F animals with early-and late-stage GC, we were able to identify proteins that were unique to each disease stage and eliminate proteins that may have been associated with aging rather than the presence of a cancer.
We used serial samples from the same animals which allowed for robust tracking of treatment success between the two modalities.Our previous studies have shown that IL-11 Mutein is efficacious in this GC model [15].Here, we demonstrate that a monoclonal antibody against IL-11R had comparable efficacy and safety to IL-11 Mutein-mediated inhibition of signaling, and thus enabled a head-to-head comparison of changes in the serum proteome throughout treatment between the different modalities.

Mice
All animal experiments were conducted in accordance with the Animal Ethics Committee (AEC) of the Ludwig Institute for Cancer Research (AEC11/07 approved 2011) and the Walter and Eliza Hall Institute, Australia (AEC 2012.019approved 2012).Age-and gendermatched gp130 Y757F mice or wild-type mice on a mixed 129/C57BL/6 background [11] were maintained in a specific pathogen-free barrier facility.
For therapeutic studies, 6.5-9.5-week-oldgender-matched gp130 Y757F mice received intraperitoneal injections of 50 mg/kg recombinant human IL-11 Mutein-PEG [15] three or 15 mg/kg of a murine specific anti-IL-11R monoclonal antibody (CSL-4E5 mG1k) [20] once/week over 4 weeks.Control animals received 15 mg/kg of an isotype control (CSL-3M4 mG1k).Sub-mandible bleeds were performed before the first treatment and at the end of the second week of treatment.Terminal cardiac bleeds were performed at the end of the treatment protocol.Serum was collected in microvette 500 Z-gel serum tubes and stored at −80 • C.
On autopsy, stomachs were excised, cut along the greater curvature and pinned with the lumen open.Macroscopic gastric tumors were enumerated according to their size and wet mass and representative polyps and adjacent (non-cancerous) antral tissue was prepared for histological analysis.Comparisons between values from two groups were performed using Student's t-test (two tailed).* p < 0.05.** p < 0.01.*** p < 0.001.

Gene Expression
For isolation of gastric cell populations, tissue was digested in 5 mM EDTA, RNA was extracted using TRIzol reagent (Invitrogen, Waltham, MA, USA), and cDNA was prepared using a High Fidelity cDNA synthesis kit (Applied Biosystems, Waltham, MA, USA) as per the manufacturer's instructions.Quantitative RT-PCR gene expression analysis was performed in triplicate using taqman probes (Applied Biosystems) on an ABI7300.Expression data were normalized to Gapdh mRNA expression.

Proteomics 4.3.1. Immunoaffinity Chromatography
High-abundance proteins were depleted from the gp130 Y757 mouse serum using the 4.6 × 50 mm Multiple Affinity Removal System ™ (MARS) column MS-3 (Agilent Technologies, Santa Clara, CA, USA) which specifically removes albumin, IgG, and transferrin.Depletion was performed at room temperature with an Agilent 1100 series HPLC system as per manufacturer's instructions.Serum samples were diluted four-fold in MARS buffer A and centrifuged at 13,000× g for 5 min to remove particulates.Low-abundance protein fractions (representing depleted serum) were collected and stored at −20 • C.

Proteomics Sample Preparation
Serum obtained from mice within the therapeutic studies were diluted four-fold in Tris/Urea buffer (0.1 M Tris-HCl, 8 M urea, pH 8.5) and 5 µL diluted serum from each mouse was prepared for mass spectrometry analysis using the FASP digestion method as previously described [48].Depleted serum fractions (400 µL) from gp130 Y757 mice were also subjected to FASP digestion.

High pH Peptide Fractionation
Tryptic peptides (3 µL) from each of the gp130 Y757 mice were pooled across the samples and subjected to high pH reversed-phase analysis on an Agilent 1100 Series HPLC system equipped with a variable wavelength detector (280 nm).Fractionation was performed on XBridge™ Shield C 18 column (10 × 100 mm, 3.5 µm bead size, Waters, Milford, MA, USA).Peptides were separated by their hydrophobicity at a high pH at a flow rate of 0.1 mL/min using a gradient of mobile phase A (5 mM ammonium formate, pH 10) and a mobile phase B (100% acetonitrile, ACN), from 3% to 45% over 70 min.Fractions were collected every minute across the gradient length and concatenated into 27 fractions.Eluted peptides were dried in a SpeedVac centrifuge (CentriVap, Labconco, Kansas City, MO, USA) and reconstituted in MS loading buffer (2% ACN/0.1% FA) prior to MS analysis.

Mass Spectrometry Experimental Design
Peptide mixtures (1 µL of samples, 2 µL for fractions) were separated by reverse-phase chromatography on a C18 fused silica column (I.D. 75 µm, O.D. 360 µm × 25 cm length) packed into an emitter tip (IonOpticks, Melbourne, VIC, Australia), using a nano-flow HPLC (M-class, Waters).The HPLC was coupled to an Impact II UHR-QqTOF mass spectrometer Bremen, Germany) using a CaptiveSpray source and nanoBooster at 0.20 Bar using acetonitrile.Peptides were loaded directly onto the column at a constant flow rate of 400 nL/min with 0.1% formic acid in MilliQ water and eluted with a 90 min linear gradient from 2 to 34% buffer B (99.9% acetonitrile and 0.1% formic acid).Mass spectra were acquired in a data-dependent manner including an automatic switch between MS and MS/MS scans using a 1.5 s duty cycle and 4 Hz MS1 spectra rate, followed by MS/MS scans at 8-20 Hz dependent on precursor intensity for the remainder of the cycle.MS spectra were acquired between a mass range of 200-2000 m/z.Peptide fragmentation was performed using collision-induced dissociation.Raw files consisting of high-resolution MS/MS spectra from the Bruker Impact II instrument were processed with MaxQuant (version 1.5.8.30) for feature detection and protein identification using the Andromeda search engine [49,50].Extracted peak lists were searched against the Mus musculus (UniProt, October 2016) databases as well as a separate reverse decoy database to empirically assess the false discovery rate (FDR) using strict trypsin specificity allowing up to two missed cleavages.In the main search, precursor mass tolerance was 0.006 Da and fragment mass tolerance was 40 ppm.The minimum required peptide length was set to seven amino acids.Modifications: Carbamidomethylation of Cys was set as a fixed modification, while N-acetylation of proteins and oxidation of Met were set as variable modifications.LFQ quantification was selected, with a minimum ratio count of 2. PSM and protein identifications were filtered using a target-decoy approach at an FDR of 1%.The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [51] partner repository with the dataset identifier PXD043224.

Data Processing and Statistical Analysis
Data processing and analysis were performed using R (version 4.1.0).The false hits, including contaminants, reverse proteins, and proteins identified by site were removed.Only proteins that were quantified in at least 50% of replicates in at least one condition were kept.The protein intensities were log 2 -transformed.Missing values were imputed by using Missing Not At Random (MNAR) method.This was achieved by substituting 'NAs' with numbers that were drawn from a normal distribution with a mean that was left-shifted from the sample mean by 1.8 standard deviation with a width of 0.3 [52].

Differential Expression and Enrichment Analysis
Data were normalized using RUVIIIC [53].The optimum k value used to remove the unwanted variation was determined based on PCA, RLE, and p-value distribution plots.The R-package limma (v.3.48.3)was used to perform the differential analysis [54].A protein was determined to be significantly differentially expressed if the false discovery rate (FDR) was ≤0.05.R-packages ggplot2 (v.3.3.5)and pheatmap (v.1.0.12)were used to visualize the results.GO enrichment analysis of the significant gene set was performed using clusterProfiler (v.4.1.3)R-package.A GO term was determined to be significant if the false discovery was ≤0.05.

Conclusions and Future Prospects
Taken together, our results highlight the potential to identify reliable biomarkers to track IGC tumor regression in response to treatment.Future studies utilizing large biobanks of human IGC samples, taking into account genetic ancestry, Helicobacter pylori status, and other clinical risk factors, will enable validation of the biomarkers that we have identified and inform selection of the most robust marker to be developed as potential diagnostic.
Author Contributions: L.F.D. performed experiments, interpreted data, prepared the results, and assisted with manuscript writing.A.P. and A.L. performed experiments.J.Y. analyzed data and prepared figures.A.I.W. and M.E.supervised components of the study.T.L.P. supervised the study,

Figure 1 .
Figure 1.Identification of early-or late-stage GC serum biomarkers.(A) gp130 Y757F mice with early GC (2 months of age, N = 5 mice per cohort) and gp130 Y757F mice with late GC (6 months of age, N = 3-5 mice per cohort).Scale bar = 6mm.Red arrow = early collection time-point, blue arrow = late collection time-point, black arrows = macroscopic tumor; (B) principal component analysis (PCA) plot of samples colored by different groups illustrating a separation of the samples based on the first principal components; volcano plots illustrating the log2 protein ratios comparing (C) wild-type versus gp130 Y757F mice with early GC; (D) wild-type versus gp130 Y757F mice with late GC; (E) gp130 Y757F mice (early versus late GC); (F) wild-type mice (early versus late).Proteins were deemed differentially regulated if the log2 fold change in protein expression was ≥1-fold and exhibited an adjusted p-value ≤ 0.05.

Figure 1 .
Figure 1.Identification of early-or late-stage GC serum biomarkers.(A) gp130 Y757F mice with early GC (2 months of age, N = 5 mice per cohort) and gp130 Y757F mice with late GC (6 months of age, N = 3-5 mice per cohort).Scale bar = 6 mm.Red arrow = early collection time-point, blue arrow = late collection time-point, black arrows = macroscopic tumor; (B) principal component analysis (PCA) plot of samples colored by different groups illustrating a separation of the samples based on the first principal components; volcano plots illustrating the log 2 protein ratios comparing (C) wildtype versus gp130 Y757F mice with early GC; (D) wild-type versus gp130 Y757F mice with late GC; (E) gp130 Y757F mice (early versus late GC); (F) wild-type mice (early versus late).Proteins were deemed differentially regulated if the log2 fold change in protein expression was ≥1-fold and exhibited an adjusted p-value ≤ 0.05.

Figure 4 .
Figure 4. Serum biomarkers following therapeutic inhibition of IL-11 signaling.(A) Principal component analysis (PCA) plot of samples colored by different groups illustrating a separation of the samples based on the first principal components.Volcano plots illustrating the log2 protein ratios comparing pre-treatment versus end of treatment between (B) control-treated mice and anti-IL-11R mAb-treated mice and (C) IL-11 Mutein-treated mice and (D) pre-treatment versus mid-treatment IL-11 Mutein-treated mice.Proteins were deemed differentially regulated when the log2 fold change in protein expression was ≥1-fold and exhibited an adjusted p-value ≤ 0.05.Heatmaps displaying the log intensities of the most differentially expressed proteins across the pairwise comparisons, including pre-treatment versus end of treatment between (E) control-treated mice and anti-IL-11R mAb-treated mice and (F) Mutein-treated mice.

Figure 4 .
Figure 4. Serum biomarkers following therapeutic inhibition of IL-11 signaling.(A) Principal component analysis (PCA) plot of samples colored by different groups illustrating a separation of the samples based on the first principal components.Volcano plots illustrating the log2 protein ratios comparing pre-treatment versus end of treatment between (B) control-treated mice and anti-IL-11R mAb-treated mice and (C) IL-11 Mutein-treated mice and (D) pre-treatment versus mid-treatment IL-11 Mutein-treated mice.Proteins were deemed differentially regulated when the log2 fold change in protein expression was ≥1-fold and exhibited an adjusted p-value ≤ 0.05.Heatmaps displaying the log intensities of the most differentially expressed proteins across the pairwise comparisons, including pre-treatment versus end of treatment between (E) control-treated mice and anti-IL-11R mAb-treated mice and (F) Mutein-treated mice.

14 Figure 5 .
Figure 5. Summary of the serum proteome changes observed during IGC progression and tumor regression.

Figure 5 .
Figure 5. Summary of the serum proteome changes observed during IGC progression and tumor regression.