Proteomics and Lipidomics in Inflammatory Bowel Disease Research: From Mechanistic Insights to Biomarker Identification.

Inflammatory bowel disease (IBD) represents a group of progressive disorders characterized by recurrent chronic inflammation of the gut. Ulcerative colitis and Crohn's disease are the major manifestations of IBD. While our understanding of IBD has progressed in recent years, its etiology is far from being fully understood, resulting in suboptimal treatment options. Complementing other biological endpoints, bioanalytical "omics" methods that quantify many biomolecules simultaneously have great potential in the dissection of the complex pathogenesis of IBD. In this review, we focus on the rapidly evolving proteomics and lipidomics technologies and their broad applicability to IBD studies; these range from investigations of immune-regulatory mechanisms and biomarker discovery to studies dissecting host⁻microbiome interactions and the role of intestinal epithelial cells. Future studies can leverage recent advances, including improved analytical methodologies, additional relevant sample types, and integrative multi-omics analyses. Proteomics and lipidomics could effectively accelerate the development of novel targeted treatments and the discovery of complementary biomarkers, enabling continuous monitoring of the treatment response of individual patients; this may allow further refinement of treatment and, ultimately, facilitate a personalized medicine approach to IBD.

1. Introduction

Inflammatory bowel disease (IBD) represents a group of chronic intestinal disorders that is characterized by recurrent inflammation affecting the gastrointestinal tract. Ulcerative colitis (UC) and Crohn′s disease (CD) are the two main clinically defined manifestations of IBD, each with distinctive clinical and pathological features (Figure 1) [1].

Figure 1

Pathomechanisms and selected biomarkers of IBD. The main pathogenesis-associated changes are shown for UC and CD. For both diseases, changes in gut microbiota, disruption of the epithelial barrier function, and chronic immune-activation are observed. Differences have also been reported, such as for immune-regulatory processes and epithelial effects. Biomarkers can be measured directly in tissue biopsies or upon release into the gut (e.g., fecal calprotectin) or the blood (e.g., autoreactive antibodies).

UC is a chronic, non-transmural, inflammatory disease that is characterized by diffuse mucosal inflammation involving the colon. UC lesions generally begin in the rectum and extend proximally in an uninterrupted pattern, involving part or all of the colon [2]. The main clinical symptom of UC is bloody diarrhea; extra-intestinal manifestations are common, including musculoskeletal, dermatological, ocular, and hepatobiliary co-morbidities. The clinical progression of UC is characterized by exacerbations and remissions [3,4,5]. From an immunological standpoint, UC present with an atypical T helper cell (Th) 2 response involvement, mediated by secretion of interleukin (IL)-13 by natural killer T cells [6,7,8].

CD is a relapsing, transmural, inflammatory disease that may affect the entire gastrointestinal tract, from the oral cavity to the rectum. It is characterized by a non-contiguous inflammatory pattern with areas of uninflamed mucosa. Its primary clinical symptom is abdominal pain with weight loss; it can also present with extra-intestinal co-morbidities similar to those described in UC patients. T cell profiles in CD and UC differ, and Th1 cytokine profiles are dominant in CD [3,4,9,10,11,12].

With its intrinsic multifactorial etiology, IBD springs from an altered interaction between the resident intestinal flora and the immune mucosal cells, due to the influence of environmental factors affecting genetically susceptible hosts. This generates a vicious cycle in which the intestinal epithelium loses its inherent integrity and becomes more pervious, priming a positive feedback loop that involves increased exposure to the intestinal microbiota, thereby leading to an uncontrolled inflammatory response [13,14,15,16].

The entangled nature of all factors contributing to its pathogenesis and their reciprocal influence led to the conceptual framing of these factors as the IBD interactome, which encompasses aspects of the immunome, microbiome, exposome, and genome [17,18,19,20,21]. In the gut microenvironment, where the multitude of IBD components functionally interact, no information about a single molecule, single gene, or single microbe can sufficiently explain the collective events that result from constant feedback and feed-forward signaling. For this reason, along with the broad variability of biological signatures in humans, an inclusive comprehension of how every component of the IBD interactome acts on and influences other components is imperative.

Unravelling the connection between known and unknown paths is needed to untangle all levels of the IBD interactome, and omics approaches—interpreted within systems biology analysis frameworks—could greatly accelerate this mission. Omics techniques allow the study of related sets of biological molecules in a comprehensive fashion; the development and use of these techniques have been rising sharply in recent decades [22]. Omics techniques include genomics, transcriptomics, metabolomics, proteomics, and lipidomics. While genetic and transcriptomic studies have been explored thoroughly in the context of IBD, proteomics and lipidomics research is now emerging as a new frontier to investigate these diseases.

In addition to yielding insights into complex etiologies, omics techniques are commonly applied for the discovery of biomarkers to support diagnosis, stratification, and treatment monitoring. For instance, although endoscopy remains the gold standard for the diagnosis and monitoring of IBD, the use of molecular biomarkers in clinical practice, including fecal calprotectin [23], serum C-reactive protein (CRP) [24], and serum autoantibodies, has been extremely important and supportive. However, their low sensitivity and high variability characteristics limit clinical efficacy [25,26]. Thus, IBD clinical management would clearly benefit from the identification of novel molecular biomarkers [27,28], which could provide less invasive assessment methods, facilitate novel targeted treatments, and decrease the medical financial burden [24,29,30].

In this review, we focus on the application of omics technologies, particularly proteomics and lipidomics, for IBD research. Driven by recent advances in mass spectrometry (MS) instrumentation, both omics approaches are rapidly evolving and could provide deeper insights into the complex pathogenesis of IBD [31,32].

2. IBD Diagnosis and Treatment: Need for Novel Approaches

CD or UC diagnosis integrates disease symptoms with data from endoscopic and histological evaluations. Treatment choices are often driven by symptom severity and patient responsiveness in order to induce remission and prevent occurrence of flares. Available IBD treatments include mesalazine, corticosteroids, immunosuppressive drugs, and monoclonal antibodies against tumor necrosis factor alpha (TNF-α) [33,34]. However, approximately one-third of patients do not exhibit improvement after induction therapy (primary non-response); loss of response may occur gradually in up to 20% of patients per year [35,36], and selection of alternative treatment strategies in non-responsive patients is extremely challenging. Importantly, treatment success depends on several patient-specific factors, including the optimization of therapeutic dosage, co-morbidities, and use of the correct disease activity index [37]. In the future, it is expected that—based on a deeper mechanistic understanding of the disease—novel molecularly targeted therapies, embedded in adaptive personalized medicine treatment frameworks, are likely to bring further therapeutic benefits to the field [38,39]. Clearly, such personalized medicine approaches will rely strongly on the availability of effective molecular companion diagnostics.

Classical diagnostic tools provide a snapshot of a few aspects of a very complex picture; the low sensitivity and variability of currently employed biomarkers, such as fecal calprotectin and CRP, limit their adoption in the clinic [25,26]. The current patent landscape for IBD biomarkers illustrates the current state of the field (Figure 2). The published patent families covering IBD biomarkers emphasize their diagnostic relevance for disease type identification (e.g., CD versus UC), determination of genetic predisposition, determination of inflammation/disease activity, and as companion diagnostics (Figure 2A). Notably, the top 20 biomarkers (on the basis of the number of patents) demonstrate the emphasis on a few common target categories; in addition to fecal calprotectin, these include anti–Saccharomyces cerevisiae antibodies (ASCA) and perinuclear anti-neutrophil cytoplasmic antibodies (p-ANCA) (Figure 2B). Especially relevant in terms of novel targeted treatment approaches, several companion diagnostics have been patented, including those for anti-TNF therapy and the specific anti-TNF drugs infliximab, adalimumab, and certolizumab (Figure 2C) [40].

Figure 2

Patent survey for IBD biomarkers. (A) Development of IBD biomarker patents over the last two decades. A worldwide IBD patent search for the last 20 years was conducted by using the Questel Orbit database. Cumulative numbers of patent families for CD, UC, and general IBD over time are shown in the line chart. The donut chart shows the current breakdown of number of patents by biomarker type. Note that a patent family can be associated with more than one disease and biomarker type. (B) Top 20 IBD biomarker patent families. The color code shows the commonly associated disease pattern for each biomarker. Note that some biomarkers appear in different variations (e.g., ANCA (PANCA, PERINUCLEAR ANTI-NEUTROPHIL CYTOPLASMIC ANTIBODY; ANCA, ANTI-NEUTROPHIL CYTOPLASMIC ANTIBODY) and ASCA (ASCA, ANTI-SACCHAROMYCES CEREVISIAE ANTIBODY; ASCA-A, ANTI-SACCHAROMYCES CEREVISIAE ANTIBODY-IMMUNOGLOBULIN; ASCA-G, ANTI-SACCHAROMYCES CEREVISIAE ANTIBODY-IMMUNOGLOBULIN G)). (C) Number of patent families disclosing companion diagnostics biomarkers for IBD treatments. (D) Registered IBD clinical trials mentioning the respective endpoints (data regarding clinical trials on IBD downloaded from ClinicalTrials.gov on 15 December 2017; text search for different endpoints in trial title, description, and endpoint fields; trials classified as UC (512), CD (695), or both (191) were considered).

Notably, the limited use of molecular biomarkers for diagnosis, stratification, and monitoring of IBD is also apparent in the available clinical trials data (Figure 2D). Complementing endoscopy and common symptom-quantifying research tools (e.g., the Crohn’s Disease Activity Index (CDAI) and Mayo score), calprotectin, and CRP are commonly used as molecular biomarkers in both CD and UC clinical trials. However, other evaluated biomarkers (ASCA/ANCA, lactoferrin, and Neutrophil gelatinase-associated lipocalin (NGAL)/Lipocalin 2 (LCN2)) have been limited to sporadic inclusion in these trials.

Overall, despite the use of some molecular biomarkers in clinical practice, there is a clear gap in terms of translating biomarker discoveries into clinical application [41]. This can be partially explained by the time required to translate omics discoveries from bench to bedside; it is also related to the complexity of designing reliable cross-border methods to ensure consistent results. Resolving this challenge and implementing individual omics profiling could enable a deeper understanding of the different molecular and clinical subtypes, thus facilitating easier and more precise diagnosis and personalized IBD treatment [42,43,44].

3. Applying Omics Analyses for IBD Research

Systems-level insights into disease mechanisms, the discovery of complex biomarkers, and, eventually, personal omics profiling are enabled by rapid advancements in molecular measurement technologies. Notably, novel sequencing technologies now facilitate the analysis of genome variations and transcriptome responses at a depth that was inconceivable just a few years ago [45]. In the IBD context, these novel sequencing methods have already provided us with novel insights, such as candidate biomarkers for CD [46]. However, as discussed, any complex biological disease, such as IBD, affects the system simultaneously on multiple levels; thus, multiple complementary analysis approaches are required to unravel relevant pathomechanisms and to identify robust biomarkers.

Here, we focus specifically on the question of how comprehensive analyses of proteome and lipidome alterations can effectively contribute to IBD research. Proteomics and lipidomics occupy different positions in the spectrum of omics methods: proteomics is the more mature and readily available method, whereas lipidomics allows completely new insights into the structure–function links of this important class of molecules. Notably, both omics methodologies, driven especially by improved MS instrumentation, are evolving rapidly, preparing the ground for future discoveries. In the following sections, we briefly summarize the contributions of these omics technologies to the investigation of IBD pathomechanisms and the identification of IBD biomarkers, before highlighting key considerations and possible improvements for future studies. As technical background, we provide a concise technical summary of current proteomics and lipidomics methods in Figure 3 and the Appendix A, with additional details available in recent review articles [31,47,48,49,50].

Figure 3

Schematic overview of the main proteomics and lipidomics technologies. (A) Gel-based and gel-free proteomics analysis workflows. In two-dimensional polyacrylamide gel electrophoresis (2D-PAGE), proteins are separated in two dimensions by molecular weight and isoelectric point. Proteins are then identified by MS after in-gel digestion. In LC-MS/MS approaches, proteins are digested into peptides prior to chromatographic separation and identification/quantification by MS. (B) Different MS-based proteomics approaches rely on distinct combinations of mass analyzers, such as quadrupoles (Q1/Q3) and high-resolution Orbitrap analyzers. To support identification, peptides can be fragmented in dedicated quadrupoles (q2) or higher-energy C-trap dissociation (HCD) cells. Untargeted proteomics involves two levels of mass spectra: the first (MS1) spectrum represents the masses of the unfragmented peptide ions. From this spectrum, peptide ions are selected for fragmentation in the HCD cell before analysis in the Orbitrap; the fragment ions are then represented in the second (MS2) spectrum. For targeted analyses, multiple reaction monitoring (MRM) commonly relies on a triple-quadrupole MS instrument. Specific peptide/fragment mass pairs (transitions) are selected and generated with quadrupole mass filters (Q1–Q3). During a targeted experiment, the mass spectrometer can cycle though several transitions to allow for multiplexing. Parallel reaction monitoring is a related technology that relies on a high-resolution fragment mass analyzer, such as an Orbitrap, rather than a quadrupole, allowing for simultaneous quantification of all fragment ions. In data-independent acquisition, ions are selected in bins over the entire analyzed mass range (Q1), all ions that fall into a mass bin are fragmented together (HCD), and the combined fragmentation spectra are deconvoluted during data processing. (C) Lipidomics can be divided into three main technologies, known as targeted, shotgun, and imaging lipidomics. Targeted and shotgun lipidomics are based, respectively, on the analysis of extracted samples, either by applying chromatographic separation or directly without separation before MS. Targeted lipidomics is performed primarily on instruments providing high sensitivity, such as triple-quadrupole MS, whereas shotgun lipidomics is preferably, but not exclusively, performed on high-resolution instruments, such as Orbitrap MS. Imaging lipidomics is based on the analysis of tissue slices through ionization of their surface molecules. The measured lipid ions are recorded by the mass spectrometer, then identified and quantified via dedicated software. For imaging lipidomics, molecules are localized onto the analyzed tissue. Figure elements based on Titz et al. [48]. See Appendix A for more details.

3.1. Mechanistic Insights Gained Using Proteomics and Lipidomics

Proteomics and, to a lesser extent, lipidomics have already been used successfully to investigate IBD pathomechanisms, including the inflammatory response, epithelial barrier function, and gut microbiome (Table S1).

The molecular coverage and quantification accuracy of current proteomics and lipidomics approaches now enable investigation of the complex immune response directly in IBD patient samples. For example, a recent proteomics investigation further elucidated the innate immune response in IBD [51]: 5711 proteins were identified in mucosal colon biopsies without visible surface inflammation. In total, 46 proteins differed in abundance between UC and control colon tissue; UC tissue exhibited clear enrichment of upregulated neutrophil and neutrophil extracellular trap proteins [52]. This identification of clear signs of chronic inflammation, even in the absence of visible surface inflammation, highlighted the need for further studies into the chronic inflammation state in IBD patients [53].

For lipidomics, a recent case–control study demonstrated how this omics method can be applied to investigate disease states in mucosa samples [54]: inflamed mucosa showed increased levels of seven eicosanoids (prostaglandin (PG) E2, PGD2, thromboxane B2, 5-hydroxyeicosatetraenoic acid (HETE), 11-HETE, 12-HETE, and 15-HETE) that correlated with the degree of inflammation [54].

Both proteomics and lipidomics have also been employed to investigate IBD-relevant immune-cell responses. Studies have implicated Th17 T-cells in IBD pathogenesis, particularly in CD-associated dysregulation of immune responses [55,56]. Riaz et al. compared the proteomes of human Th1 and Th1/Th17 clones derived from intestinal biopsies of CD patients [57]. In total, 7401 proteins have been quantified; 334 were differentially expressed between Th1 and Th1/Th17 clones. Consistent with their functions in immune responses, cytotoxic proteins, such as granzyme B and perforin, were more abundant in Th1 than in Th17 cells. However, only a subgroup of Th1 cell clones from CD patients, characterized as CD28-positive and natural killer group 2 member D-negative, expressed these cytotoxic features; this suggested a larger-than-expected diversity in the T cell-mediated immune response in CD. In another recent example, proteomics of regulatory T-cells (CD4+Foxp3+) (Treg) led to the identification of a protein, THEMIS, as a checkpoint control in the suppressive function of Treg cells [58]. In a related context, lipidomics pinpointed specific differences in the response of macrophages from CD patients upon activation with heat-inactivated Escherichia coli, with significantly lower levels of newly synthesized phosphatidylinositol 16:0–18:1 [59].

As discussed earlier in this review, the disruption of intestinal epithelium integrity is an early event in IBD pathogenesis. Proteomics has been used to study differences between intestinal epithelial cells isolated from IBD patients and controls, leading to the identification of activated cellular stress responses in IBD [60,61].

Given the relevance of intestinal mucus as an intestinal barrier, lipidomics was used to investigate the phosphatidylcholine (PC) lipidome profile of rectal mucus obtained from IBD patients and control subjects [62,63]. UC patients displayed significantly lower levels of PC and lyso-PC compared with CD patients and controls. Interestingly, treatment of UC patients with a special formulation of PC, which exhibited delayed release in the gut, showed clinical efficacy [64], resulting in an improvement of the clinical activity index [65].

Genomics methods are at the forefront of the dissection of gut microbial populations. Nevertheless, metaproteomics is emerging as an effective approach to gain further insights into the complex interactions between the gut microbiome and host environment in health and disease [66]. In a recent study, a systems biology approach, including proteomics, was used to further understand the host–microbe cross-talk in new-onset CD [67]. The authors analyzed microbiota of the mucosa-luminal interface by 16S rDNA gene sequencing and combined this with a quantitative proteomics investigation of the host proteomes in mucosal biopsies; 320 of 3323 quantified proteins were significantly differentially expressed, including several mitochondrial proteins that were significantly downregulated in CD patients compared with controls. Interestingly, this included several proteins involved in H2S detoxification. Concomitantly, metagenomics data showed an increase in the relative abundance of microbial H2S producers, indicating Atopobium parvulum as a central hub. Finally, supporting a possible causal link, in a follow-up experiment with IL-10-deficient mice, A. parvulum was shown to induce colitis, which was mitigated by an H2S scavenger [67].

3.2. Biomarker Identification Using Proteomics and Lipidomics

In addition to these mechanistic studies, proteomics and lipidomics have been applied in several IBD biomarker identification studies (Table S1). In total, we identified 18 published studies that were performed specifically to identify biomarkers. Only a single study employed lipidomics, and most recent proteomics studies tended to employ liquid chromatography coupled to tandem MS (LC-MS/MS) rather than 2D-PAGE technologies, reflecting a shift to more recent and more potent proteomics technologies. Most of the studies (11 in total) focused on blood/serum as the sample source; five studies focused on colon samples, two studies on feces, and a single study on saliva, aligned with the convenience of sampling (serum, feces, and saliva) and proximity of the sampling site to the disease process (colon biopsies and feces). These studies were undertaken to identify biomarkers in at least three different application areas: diagnosis, patient stratification, and treatment selection and response monitoring.

For example, a recent study, which was focused on IBD diagnosis, combined discovery proteomics with targeted verification experiments to identify markers for intestinal complications of CD (untreated/balloon-dilated strictures and non-healed abscesses/fistulas) [68]. The derived serology panel was able to indicate complications in CD with 70% sensitivity and 72.5% specificity. In a lipidomics study, plasma lipid profiles were compared across healthy controls and UC and CD cases [69]; this resulted in the identification of 33 lipid species, primarily belonging to ether-lipids, that were significantly correlated with CD.

Another recent proteomics study focused on proteins at the mucosal-luminal interface to identify biomarker candidates for pediatric IBD-associated colitis [70]. Two distinct four-protein panels were identified that were able to discriminate active IBD from non-IBD and pancolitis from non-pancolitis, with reported high sensitivity and specificity (>0.95). However, these estimates were obtained from the discovery cohort, not from an independent validation cohort; thus, it is unclear how these panels might perform with real clinical samples. Nevertheless, one protein from each panel, LTA4H, and catalase, was also measured in stool samples from an independent cohort, demonstrating elevation in IBD versus non-IBD stool samples. These data support the selection of IBD-relevant proteins and demonstrate the possibility that these markers can be measured in more easily accessible stool samples.

The identification of specific IBD phenotypes in patients diagnosed with indeterminate colitis (IC) has been implemented recently by using disease stratification biomarkers. Among young (<18-year-old) IBD patients, 15−30% cannot be fully classified and are therefore diagnosed with IC. While individual serum cytokine profiles are unable to distinguish among IBD subtypes [71,72], both transcriptomics [71] and MALDI imaging proteomics profiles of colon biopsies [73,74] showed clear differences between UC and CD, supporting the feasibility of the development of biopsy biomarkers for disease stratification.

Another study, intended to improve disease stratification, identified protein markers that were able to discriminate between CD and UC in children with new-onset IBD [75]. Two candidate biomarker panels were established: a five-protein panel able to discriminate IBD from control cases with 95.9% accuracy, and a 12-protein panel able to discriminate between CD and UC patients with 80% accuracy in the validation cohort.

Proteomics has also been employed to identify treatment response biomarkers for IBD. Anti-TNF therapy with infliximab is now commonly used to control inflammation in IBD patients [76]. However, the factors predicting response, as well as the cellular mechanisms leading to the therapeutic response, are not fully understood. In a population of 43 UC patients, Magnusson et al. investigated infliximab treatment response in blood and inflamed colon biopsies [77]. Treatment response was associated with lower levels of macrophage markers CD14 and CD86 and the chemokine CCL2. In addition, proteomics identified reduced levels of tenascin-C (a glycoprotein of the extracellular matrix) in biopsies upon treatment, which was consistent with reduced tenascin-C levels in serum.

Finally, another study investigated differences in the serum protein profiles of children with IBD before and after treatment with infliximab or corticosteroids [78]. In total, 18 proteins differed significantly in both treatment groups with the same directionality; consistently downregulated proteins were associated with inflammatory processes.

4. Proteomics and Lipidomics in IBD Research—Going Forward

As illustrated in the above sections, both proteomics and lipidomics can provide further insights into IBD pathomechanisms and support the development of novel biomarker assays. However, thus far, both these omics technologies have not exploited their full potential. The published studies provide a fairly heterogeneous picture of the field, with interesting case studies; however, there is a lack of clear direction regarding how these findings can be further developed and, eventually, translated into clinical practice. This is especially evident in the biomarker discovery field, where only a few studies have included a validation cohort, and there is an obvious disparity between reported biomarker candidates and those used in the clinic. Clearly, translating novel findings is a long process. Nevertheless, design of new studies can clearly benefit from the newly generated data taking into consideration some key aspects (Figure 4).

Figure 4

Aspects to consider in the design and conduct of IBD biomarker studies.

Omics technologies. MS-based technologies for the quantification of proteome and lipidome changes are evolving rapidly (see Appendix A). The molecular coverage and sample throughput available today were unimaginable just a few years ago. Noteworthy, for example, is the emergence of so-called data-independent acquisition approaches, which now enable very comprehensive and reproducible snapshots of the proteome in large sample cohorts [79]. This growth will likely continue, and future studies will benefit from advances in omics technologies.

Sample quality. However, it is important not only to focus on the specific omics technology employed in a given situation but also to design the entire sample collection, processing, and analysis workflow from a robustness and quality perspective. Sample quality is among the most important aspects required for a successful biomarker study [80]. Well-defined standard operating procedures for sample collection and handling can ensure high-quality, reproducible sampling [81], as reported for serum/plasma [82] and cerebrospinal fluid [83].

Analytical platform. To ensure the overall quality of the analytical platform, its performance must be characterized in depth (e.g., in terms of measurement accuracy and reproducibility). For proteomic studies, the scientific community has developed and published useful guidelines [84,85,86]. Rigorous analytical quality control is crucial and can be supported by automated LC-MS/MS performance monitoring systems [87,88], together with specific sample quality metrics (e.g., [89,90]). To prevent late failures of biomarker candidates, it has been suggested that a core method validation (e.g., including estimates of accuracy, precision, and limits of quantification) should be performed when designing biomarker discovery workflows [91]. Clearly, full method validation becomes essential when a newly discovered biomarker is moved toward clinical application. For this, recognized validation guidelines are to be followed, such as those provided by the Clinical and Laboratory Standards Institute (e.g., [92]) and those recognized by the U.S. Food and Drug Administration [93]. The Eurachem guide by Magnusson and Örnemark is a useful reference for general analytical method validation guidelines [94]. Moreover, especially for large-scale studies, automation can be considered to make biomarker discovery platforms more robust, as recently demonstrated by several groups (Table 1).

Table 1

Recently published LC-MS/MS-based protein biomarker identification workflows.

Quantification Approach	Abundant Protein Depletion	Separation	Scale(#Proteins/#Samples)	Precision	Comment	References
iTRAQ (discovery)	Yes	SCXC18	Pool of 20 samples		Biomarker panel for diabetic kidney disease developed	[95]
MRM (validation)	Yes	C18	8/572	CV ~8%
Label-free	No	C18	~437/319	CV <20% for 71% of proteins	Applied within weight-loss study	[96]
TMT	Yes	C18	~190/~1000	Average CV of 12% for internal standard	Applied within multicentered dietary interventionstudy	[97]

Coefficient of variation (CV); strong cation exchange chromatography (SCX); octadecyl carbon chain chromatography (C18).

Study design. Ultimately, especially for biomarker studies, the study design will determine the success of a project. Relevant considerations for the design of biomarker identification and validation studies have been outlined [81,98,99]; these include clearly defined inclusion and exclusion criteria, control of confounding factors, and well-defined statistical analysis plans with statistical power calculations [100,101]. If possible, longitudinal designs (i.e., following biomarkers in a subject over time) should be considered, as these are especially powerful for assigning the observed variability in a candidate biomarker to a biological response of interest, rather than to complex differences in the human population. Most importantly, however, no biomarker identification study should be planned without an independent study cohort to validate the findings. This is especially relevant considering that only 23% of the published plasma proteomics studies for biomarker discovery have been independently verified [91].

Emerging sample types. Another generally important aspect is the selection of the biological source material. A broad spectrum of biological sample types, ranging from serum to colon biopsies and feces, has been leveraged for IBD investigations. Alternative sources could be considered to facilitate new discoveries. For example, preserved tissue samples stored in large biobanks could provide an alternative and, possibly, complement studies that use fresh tissue biopsies. Several groups have shown that archived tissue samples are suitable for proteomics studies [102,103,104,105], with a recent study applying proteomics to fixed colon mucosa samples [106]. Another promising sample type for IBD research is whole gut lavage fluid (WGLF) [107], which can be collected when the gastrointestinal tract is cleansed before colonoscopy or colon surgery. For example, in one study, lactoferrin was increased in the WGLF of IBD patients—it was thus proposed as a marker for intestinal inflammation [108]—while elevated levels of IL-1β were identified in the WGLF of CD patients with a high risk of relapse [109]. Finally, colorectal mucus could be further explored as a potential sample source. Previous studies have established that colorectal mucus is a rich source of potential IBD biomarkers, as demonstrated by the quantification of calprotectin and eosinophil-derived neurotoxin in mucus samples from IBD patients [110,111].

Multi-omics. Returning to the main theme, another key general consideration is which molecule class to profile. While we focus here on proteome and lipidome profiling methods, clearly, transcriptome, (epi-) genome, and metabolome profiling methods can be expected to yield very relevant insights. Importantly, IBD pathogenesis results from disturbances in the complex interplay of multiple biological processes; a comprehensive understanding of these processes will ultimately require integrative approaches combining several omics and non-omics endpoints within a systems biology framework [20,44,112,113]. Notably, Danese et al. recently emphasized the need for multi-omics integration to achieve comprehensive and effective IBD disease risk assessment, early and accurate diagnosis, effective targeted treatments, identification of the biological basis of disease states, and, finally, disease prevention [114]. Similarly, as mentioned, it has been argued that the overwhelming complexity of IBD can only be addressed with an integrative view of all pathological components, known as the “IBD interactome” [44]. Generating relevant and interpretable views of multi-omics data sets remains a challenge; however, several omics data integration methods are readily available [115], and novel methods are emerging (e.g., Multi-Omics Factor Analysis [116]).

The multi-level insights thus gained into IBD pathogenesis could then lead to the development of effective biomarkers that are specific for central regulatory nodes in the disease process. Importantly, how feasible are such multi-level or multi-omics biomarkers? The benefits of such integrative biomarker panels have already been recognized for IBD, with commercially available products that combine serologic, genetic, and inflammation markers to support IBD diagnosis (Table S2).

5. Conclusions

IBD is a complex disease with a multifactorial etiology that is affected by genetics, gut microbiota, immune response, and the environment [17,18,19]. Global molecular profiling and quantitative technologies, including proteomics and lipidomics—in combination with sophisticated computational analysis and integration approaches—are emerging as new tools able to generate high-throughput data, instrumental for further understanding of IBD pathogenesis and biomarker identification.

In this review, we have summarized the state-of-the-art proteomic and lipidomic approaches to IBD studies. The intrinsic conundrum of intestinal physiology and its derangement in gut pathologies put in place different players in IBD onset and development, thus creating different layers of complexity. The application of targeted omics studies will enable unraveling of the contributions of each tissue subcomponent. The rapid evolution of these technology platforms will transform the current step-up approach [117], in which treatment is chosen on the basis of disease severity and treatment responsiveness. Importantly, it will accelerate biomarker discovery, indicate novel therapeutic avenues for the clinical management of IBD, and facilitate the possibility of continuous monitoring of the treatment response of individual patients, allowing further refinement of treatment. Implementing the output of omics techniques, in combination with (epi-)genomics, microbiome studies, and nutrigenomics, will eventually lead to personalized medicine for IBD patients (Figure 5).

Figure 5

Step-up and personalized treatment approaches for IBD. In the classical “step-up” approach to IBD treatment, treatments are selected and aligned with disease severity, from the bottom to the top of the pyramid. The proteomics and lipidomics approaches discussed in this review will contribute to the development of personalized treatment of IBD patients. The discovery of predictive biomarkers will support the matching of personalized treatment, continuous monitoring of treatment responses, and further treatment refinements.