Construction of 2DE Patterns of Plasma Proteins: Aspect of Potential Tumor Markers

The use of tumor markers aids in the early detection of cancer recurrence and prognosis. There is a hope that they might also be useful in screening tests for the early detection of cancer. Here, the question of finding ideal tumor markers, which should be sensitive, specific, and reliable, is an acute issue. Human plasma is one of the most popular samples as it is commonly collected in the clinic and provides noninvasive, rapid analysis for any type of disease including cancer. Many efforts have been applied in searching for “ideal” tumor markers, digging very deep into plasma proteomes. The situation in this area can be improved in two ways—by attempting to find an ideal single tumor marker or by generating panels of different markers. In both cases, proteomics certainly plays a major role. There is a line of evidence that the most abundant, so-called “classical plasma proteins”, may be used to generate a tumor biomarker profile. To be comprehensive these profiles should have information not only about protein levels but also proteoform distribution for each protein. Initially, the profile of these proteins in norm should be generated. In our work, we collected bibliographic information about the connection of cancers with levels of “classical plasma proteins”. Additionally, we presented the proteoform profiles (2DE patterns) of these proteins in norm generated by two-dimensional electrophoresis with mass spectrometry and immunodetection. As a next step, similar profiles representing protein perturbations in plasma produced in the case of different cancers will be generated. Additionally, based on this information, different test systems can be developed.


Introduction
In a broad sense, tumor biomarkers are components that are either produced directly or indirectly because of a tumor. Moreover, these biomarkers can be common cellular products that are overproduced by cancer cells or the products of genes that are expressed only during malignant transformation. Thus, a tumor marker that is present in significant quantities indicates the presence of cancer. The marker can be present inside the tumor or enter the bloodstream [1,2]. This point is fundamentally important, as it allows the noninvasive examination and treatment of patients with various malignant neoplasms. The list of biochemical tumor markers known today is large [2]. Although some of these biomarkers have been successfully used in treatment, none of them fully satisfy the so-called "ideal marker", which should be highly sensitive, specific, reliable with high predictive value, and correlate with the stages of tumor development [3].
Therefore, the search for new markers continues. Here, multi-omics technologies such as genomics, transcriptomics, and metabolomics are very important, but proteomics plays a central role since tumor biomarkers are mostly proteins. From a proteomic point of view, the search is based on a comparative analysis of proteomes. These proteomes are from body fluids (blood plasma, cerebrospinal fluid, saliva, urine, etc.) or tissues. Here, human plasma is one of the most popular clinical samples as it provides noninvasive, rapid analysis for any type of disease. A special human plasma proteome project (HPPP) project was initiated in 2002 (https://www.hupo.org/plasma-proteome-project accessed on 10 September 2022). Now, this initiative has achieved great success in plasma protein analysis (http://plasmaproteomedatabase.org/index.html accessed on 10 September 2022) [4,5]. One of the main advantages of using plasma samples is that only a minimally invasive assay such as a routine blood test analysis is required. To the greatest extent, this certainly concerns the hematopoietic organs (for instance, the major human plasma proteins are synthesized mostly in the liver), but also applies to other tissues, and even the brain, which is separated by the blood-brain barrier. It is expected that the blood plasma proteome should reflect, to varying degrees, changes in cellular proteomes caused by diseases. In recent years, biomarker selection guidelines have been developed [6][7][8][9][10]. Here, the classical proteomic approaches are used: two-dimensional electrophoresis (2DE), immunodetection, and mass spectrometry (MS), which have many methodological options that allow highly productive analysis individually or together in different combinations. Electrophoretic separation of plasma proteins offers a valuable diagnostic tool, as well as a way to monitor clinical progress [11]. MS measures, with high accuracy, the masses of peptides obtained by specific hydrolysis of proteins and is very specific. This approach was applied for detecting ovarian cancer (OC) based on just MS-spectra [12]. In addition, MS-based proteomics can detect and quantify protein variants-proteoforms [13]. Ideally, MS-based proteomics can analyze a whole proteome [14][15][16]. A rapid, robust, and reproducible shotgun plasma proteomics workflow was developed to produce "plasma proteome profiles" [14,17].
Accordingly, there are several directions for proteomics to develop ideal oncomarkers. First, we can go deep-find highly specific proteoforms/oncomarkers secreted by a tumor in low abundancy. Second, go wide-select, and analyze a panel of multiple proteins/oncomarkers. Third, combine these approaches. There are already some examples of generation from such panels [18]. This strategy can be applied to solid or liquid biopsies depending on the real situation. Here, the question arises about how to select these oncomarkers, as the concentration range of putative oncomarkers in plasma is very wide. The plasma proteome is the most complete version of the whole human proteome. In addition to the "classical plasma proteins", it contains tissue proteins plus numerous individual immunoglobulins [19,20]. In clinics, a lot of information about the health state is obtained by analysis of blood proteins. Accordingly, in diagnosis and therapeutic monitoring, human plasma proteome analysis is a promising solution. The major protein, albumin, accounts for~50% of the mass of all proteins. Nine proteins (IgG, apolipoprotein A1, apolipoprotein A2, transferrin, fibrinogen, haptoglobin, alpha1-antitrypsin, transthyretin) make up 40%, another 12 make up the next 9%, and the rest only 1%. Accordingly, it is common practice to remove the most abundant proteins (deplete) before deep proteomics analysis of plasma [21].
Two-dimensional electrophoresis analysis of human plasma proteins has a long history, where, possibly, the input of L. Anderson and N.G. Anderson is most impressive [22][23][24]. There are many publications where the 2DE image of plasma proteins was used as a specific profile for testing the cancer-related changes in the human body [25][26][27][28][29]. However, if we are going to decipher the whole panel of plasma proteins as a combined tumor biomarker, we need to obtain reliable data about every protein in connection to its response during the malignancy process. Previously, we started to collect information about the proteoform profiles of different cellular proteins into a database "2DE pattern" using our original approaches [30]. These approaches are time consuming and labor intensive but allow the presentation of panoramic data about different proteoforms and could be very useful in biomarker studies. Here, as a next step in searching for specific oncomarkers, we produced 2DE profiles for the human plasma proteins. The most abundant, "classical plasma proteins" were selected as they are detected reliably by common proteomics methods.

Results
In our study, using classical 2DE, sectional 2DE, and semi-virtual 2DE in combination with liquid chromatography-electrospray ionization tandem mass spectrometry (LC ESI-MS/MS), we generated 2DE patterns for the most abundant plasma proteins. In Figure 1, these 2DE images of plasma proteins are presented. The 2DE patterns of more than 100 reliable and confidently detected sets (Supplementary Tables S1 and S2) are presented in Supplementary Figure S1. We also collected data from the literature about the possibilities of using these plasma proteins as cancer biomarkers (Table 1) [20,31]. The detailed information about these proteins and the 2DE patterns of plasma proteins in norm generated in our experiments are described below and in the Supplementary File.  Tables S1 and S2) are presented in Supplementary Figure S1. We also collected data from the literature about the possibilities of using these plasma proteins as cancer biomarkers (Table 1) [20,31]. The detailed information about these proteins and the 2DE patterns of plasma proteins in norm generated in our experiments are described below and in the Supplementary File.   Table S1). In the column "Cancer", the references for cancer-related data are shown (the details are in the Supplementary File).

LEUCINE-RICH ALPHA-2-GLYCOPROTEIN (A2GL_HUMAN)
The two-dimensional electrophoresis pattern of LRG1 represents a chain of spots in the pI-range 3.5-5.0 with Mw~40,000-50,000 (Supplementary Figure S1). This pattern is well-represented in the SWISS-2DPAGE https://world-2dpage.expasy.org/ accessed on 10 September 2022 and has a characteristic for multiple glycosylation profiles, where acidic spots have higher Mw [22]. LRG1 has at least six sites of glycosylation: one is O-GalNAc and five are N-GlcNAc (https://www.uniprot.org/uniprot/P02750 accessed on 10 September 2022).

ALPHA-1-ANTICHYMOTRYPSIN (AACT_HUMAN)
The two-dimensional electrophoresis pattern of ACT represents a chain of spots in the pI-range 4.0-5.0 and Mw 50-60,000 (Supplementary Figure S1). This pattern is also well-represented in the SWISS-2DPAGE, where two chains (20 spots) of both ACT forms are presented [22]. ACT has seven sites of N-GlcNAc and four sites of O-GalNAc (https://glygen.org/protein/P01011#glycosylation accessed on 10 September 2022).

ADIPONECTIN (ADIPO_HUMAN)
The two-dimensional electrophoresis pattern of adiponectin represents a chain of spots in the pI-range 5.0-5.5 and Mw 26,000 (Supplementary Figure S1). There are six sites of O-linked glycosylation and two sites of phosphorylation in adiponectin (https://glygen. org/protein/Q15848#glycosylation accessed on 10 September 2022).

AFAMIN (AFAM_HUMAN)
The two-dimensional electrophoresis pattern of adiponectin represents a chain of spots in the pI-range 4.5-6.0 and Mw~70,000 (Supplementary Figure S1). There are six sites of N-linked glycosylation in afamin, and more than 90% of the glycans are sialylated (https://glygen.org/protein/P43652#glycosylation accessed on 10 September 2022).

ALBUMIN (ALBU_HUMAN)
The two-dimensional electrophoresis pattern of albumin represents a chain of spots in the pI-range 5.5-6.5 and Mw~70,000 (Supplementary Figure S1). Its pattern is also well-represented in the SWISS-2DPAGE [22]. Albumin can be modified by N-linked glycans at one site, 7 O-linked glycans at 11 sites, phosphorylated at multiple sites (at least 15), and acetylated (1 site) (https://glygen.org/protein/P02768#glycosylation accessed on 10 September 2022).

BETA-2-GLYCOPROTEIN 1 (APOH_HUMAN)
The two-dimensional electrophoresis pattern of apo-H represents a chain of spots in the pI-range 6.2-8.4 and Mw~52,000 (Supplementary Figure S1), which is much higher than the theoretical one because of heavy glycosylation (85 N-linked annotations at 4 sites and 3 O-linked annotations at 3 sites) (https://glygen.org/protein/P02749#glycosylation accessed on 10 September 2022).

Complement System
The results of several studies suggest that changes in the complement system can not only promote an antitumor response but can also influence tumor development through proliferation, survival, angiogenesis, and invasiveness [267,268]. The presence of many complement components with different functions makes the study of this system very difficult [269]. In any case, it is becoming clear that complement activation stimulates carcinogenesis and protects against immune destruction, although it has long been believed that the complement system helps the body identify and eliminate transformed cells. Moreover, the complement is activated by different mechanisms in the case of different types of cancer, and the results of activation may be different for different types of cancer or over time for the same tumor [270][271][272].

C1R (C1R_HUMAN)
The two-dimensional electrophoresis pattern represents only a chain of spots of the complement C1r subcomponent in the pI-range 4.5-6.2 and Mw~80,000 (Supplementary Figure S1) that corresponds to only a complement C1r subcomponent. The cleaved heavy and light chains were not detected. There are 25 N-linked glycosylation annotations at four sites and one phosphorylation site in the complement C1r subcomponent (https: //glygen.org/protein/P00736#glycosylation accessed on 10 September 2022).

C1S (C1S_HUMAN)
The two-dimensional electrophoresis-pattern represents a chain of five spots of C1s in the pI-range from 4.0 to 4.9 and Mw~80,000 (Supplementary Figure S1). The cleaved heavy and light chains were not detected. There are seven N-linked glycans at two sites (https://glygen.org/protein/P09871#glycosylation accessed on 10 September 2022).

COMPLEMENT C1qC (C1QC_HUMAN)
The two-dimensional electrophoresis pattern of C1q represents a long horizontal chain of spots in the pI-range 3.0-9.5 with Mw~23,000 and a vertical chain of heavy complexes (Mw 23,000 and up) with pI~9.0 (Supplementary Figure S1). It was reported there was only one O-linked glycosylation of C1q (https://glygen.org/protein/P02747#glycosylation accessed on 10 September 2022).

COMPLEMENT FACTOR I (CFAI_HUMAN)
The two-dimensional electrophoresis pattern of the complement factor I represents the chains of many spots in the pI-range 4.5-6.8 from Mw~64,000 (complement factor I) to Mw~30,000 (the complement factor I heavy and light chains) (Supplementary Figure S1). In the SWISS-2DPAGE, the complement factor I is represented only by one spot (pI/Mw: 5.03/37,900). There are 57 N-linked glycosylation annotations at 6 sites for the complement factor I (https://glygen.org/protein/P05156#glycosylation accessed on 10 September 2022).

COMPLEMENT FACTOR B (CFAB_HUMAN)
The two-dimensional electrophoresis pattern of the complement factor B represents the chains of spots in the pI-range 4.5-6.8 with Mw~90,000 (Supplementary Figure S1). The cleaved heavy and light chains were not detected. In the SWISS-2DPAGE, the complement factor B is represented by a chain of six spots (pI 5.88-6.28, Mw~100,000). There are 19 N-linked glycans (4 sites), and 3 O-linked glycans (3 sites) in the complement factor B (https://glygen.org/protein/P00751#glycosylation accessed on 10 September 2022).

COMPLEMENT FACTOR H (CFAH_HUMAN)
The two-dimensional electrophoresis pattern of the complement factor D represents a long chain of spots in the pI-range 5.5-7 with Mw~140,000 (Supplementary Figure S1). It was reported there were 62 N-linked glycans in 9 sites in the complement factor D (https://glygen.org/protein/P08603#glycosylation accessed on 10 September 2022).

COMPLEMENT C3 (CO3_HUMAN)
The two-dimensional electrophoresis pattern of the complement C3 represents a cluster of spots in the pI-range 3.5-7.5 and Mw from~30,000 to 180,000 (Supplementary Figure S1 The two-dimensional electrophoresis pattern of the complement C4-B represents a wide cluster of spots in the pI-range 5.0-6.8 and Mw~70,000-190,000 (Supplementary Figure S1). It was reported there were eight N-linked glycans at three sites (https: //glygen.org/protein/P01031#glycosylation accessed on 10 September 2022).
2.27.14. COMPLEMENT C7 (CO7_HUMAN) The two-dimensional electrophoresis pattern of the complement C7 represents a chain of spots in the pI-range 4.5-6.5 and Mw~100,000 (Supplementary Figure S1). It was reported there were three N-linked glycans at two sites and two O-linked glycans at one site (https://glygen.org/protein/P10643#glycosylation accessed on 10 September 2022).

CARBOXYPEPTIDASE N CATALYTIC CHAIN (CBPN_HUMAN)
The two-dimensional electrophoresis pattern of CBPN represents a chain of spots in the pI-range 4.5-7 and Mw~50,000 (Supplementary Figure S1). It was reported there were 32 N-linked glycans at 5 sites and 2 O-linked glycans at 1 site https://glygen.org/protein/ P08185#glycosylation accessed on 10 September 2022.

CHOLINESTERASE (CHLE_HUMAN)
The two-dimensional electrophoresis pattern of cholinesterase represents a chain of five spots in the pI-range 4.5-5.2 and Mw~65,000 (Supplementary Figure S1). There are 34 N-linked annotations at 12 sites, one O-linked annotation for glycosylation, and phosphorylation at S226 for cholinesterase (https://glygen.org/protein/P06276#glycosylation accessed on 10 September 2022).

VITAMIN D-BINDING PROTEIN (VTDB_HUMAN)
The two-dimensional electrophoresis pattern of this protein represents a cluster of spots (pI/Mw

Discussion
Tumorigenesis leads to multiple variations in the human plasma proteome that can be dynamic and alterable during the progress of the disease. Practically all major, so-called "classical", plasma proteins change abundances or PTMs. The majority of these proteins are secreted by the liver, so it could be anticipated to see these changes only in the case of liver cancer. However, they can be observed with other tumors as well as cancer induces disturbances in the blood homeostasis that is supported by "classical plasma proteins". It follows that it is possible to search the specific/unspecific ways of tumor prediction not only through the detection of products of the tumor but also by analyzing the changes in "classic plasma proteins". It is relevant to mention that plasma analysis by a very different approach, differential scanning calorimetry (DSC), can give us a hint. Typically, DSC is used to determine the partial heat capacity of macromolecules as a function of temperature, from which their structural stability during thermal denaturation can be assessed. The method is very sensitive and allows precise determination of thermallyinduced conformational transitions of proteins present in plasma. There are already quite a few publications showing that DSC can be used to distinguish between normal and cancerous plasma samples [273,274]. Moreover, the data obtained by this method can be reproduced using major plasma proteins.
It follows that there is a possibility of building test systems based on these major ("classical") proteins. What is important is that many examples of such systems have been introduced already. For example, the relationship between inflammation and clinical outcome is described using the Modified Glasgow Prognostic Scale (mGPS), which includes levels of C-reactive protein (CRP) and albumin [275]. The combination of elevated CRP (>10 mg/L) and decreased albumin (<35 g/L) corresponds to higher mGPS, which correlates with systemic inflammation and poor outcome of cancer therapy [276]. The OVA1 test uses the other major plasma proteins. OVA1 is an FDA-approved blood test that measures the levels of five proteins (CA125, transferrin, transthyretin, apolipoprotein A1, and beta-2 microglobulin) to detect ovarian cancer risk in women. Here, a sophisticated mathematical formula (multivariate index assay) is used to evaluate and combine the levels of these proteins in plasma, producing an ovarian cancer risk score. Using this approach, OVA1 can detect early-stage ovarian cancer with 98% specificity. The OVERA (second-generation or OVA2) assesses a woman's malignancy risk using combined results from the following five proteins: apolipoprotein A1, human epididymis protein 4 (HE4), CA-125 II, folliclestimulating hormone (FSH), and transferrin (Vermillion Inc. OVA1 Products. Updated 2020. Available at: https://vermillion.com/ova-products accessed on 10 September 2022). The observation of enhanced levels of clusterin, ITIH4, antithrombin-III, and C1RL in sera of endometrial cancer patients allowed a mathematical model to be built to detect cancer samples [29]. Accordingly, by the selection of the appropriate panels (proteomics signatures) of the plasma oncomarkers, it is possible to detect/monitor different types of cancers. The main point is to select the correct set of oncomarkers and develop an algorithm that will take into account all possible changes in these oncomarkers (level, PTMs etc.) that are related to cancer. This selection should be meticulously performed based on oncomarker behavior in plasma, not in tissue. We performed a search for publications with information (level, PTMs) about "classical" plasma proteins in the case of malignant processes in the human body (Table 1). As levels of some oncomarkers behave differently in different cancers (rise or fall), the test could specifically detect the type of cancer. Apolipoproteins are a good example here. SAA1 and CRP are APPs that are routinely measured in the clinic. The level of apoA-1 is reduced in many cancers but increased in some [80]. The decreased level of apoA-I in plasma is observed in the case of de novo myelodysplastic syndromes [83], NSCLC [84], nasopharyngeal carcinoma (NPC) [85], esophageal squamous cell carcinoma [86], and BC [75], but it is increased in SCLC, HCC, and bladder cancer [80]. The level of apoA-II is dramatically reduced in the serum of patients with gastric cancer and multiple myeloma [70,87] but increased in HCC and prostate cancer [88,89]. A similar situation can be observed for other apolipoproteins [80].
Another aspect that should be considered is the appearance of proteoforms produced by genetic polymorphisms, alternative splicing, PTMs, etc. These events change the charge (pI) and the weight (Mw) of the protein. Because of that, the experimental pI/Mw of the proteins can be different from the theoretical ones. This leads to the production of sets of proteoforms that in our case are detected as 2DE patterns. There is a belief that some 2DE patterns can be different between norm and cancer and could be used as specific biomarkers. Thus far, there are not many such examples, but progress in proteomics methods should improve the situation [277,278]. Proteomics is generating and analyzing a large volume of data and these data exactly fit the situation with multiple variations in plasma proteomes during cancer development and progression. Here, high-throughput, quantitative mass spectrometry is the best choice. There is already a good example of the possibility of using it in the clinic [14]. Geyer et al. introduced a rapid and robust "plasma proteome profiling" LC-MS/MS pipeline. Their single-run shotgun proteomics workflow enables quantitative analysis of hundreds of plasma proteins from just 1 µL of plasma [14].
Our aim is to build a comprehensive proteoform database containing norm and cancer samples http://2de-pattern.pnpi.nrcki.ru/ accessed on 10 September 2022 [30]. Glioblastoma and hepatocellular carcinoma are the cancers in our study so far. The database contains only the cellular samples, but we are in the process of incorporating tissue and plasma samples.

Plasma
The pooled human plasma was from healthy male donors (age 20-47 years) [278,279]. Depletion of serum albumin and immunoglobulins IgG was carried out according to Agilent Multiple Affinity Removal System (MARS) protocol ("Agilent Technologies", Santa Clara, CA, USA) [280,281].

Two-Dimensional Electrophoresis
The detailed process was described previously [280]. In short, 10 µL of plasma (0.5 mg of protein) was mixed with 20 µL of lysis buffer (7 M urea, 2 M thiourea, 4% CHAPS, 1% DTT, 2% (v/v) ampholytes, pH 3-10, protease inhibitor cocktail) and then with 100 µL of rehydrating buffer (7 M urea, 2 M thiourea, 2% CHAPS, 0.3% DTT, 0.5% IPG (v/v) buffer, pH 3-11 NL, 0.001% bromophenol blue). Immobiline DryStrip 3-11 NL (7 cm) was passively rehydrated by this solution for 4 h at 4 • C. IEF was run on Hoefer™ IEF100 ("Thermo Fisher Scientific", Waltham, MA, USA). After IEF, strips were incubated 10 min in the equilibration solution (50 mM Tris, pH 8.8, 6 M urea, 2% SDS, 30% (v/v) glycerol, 1% DTT), following in the same solution with 5% IAM instead of DTT. The strips were sealed with a hot solution of 0.5% agarose prepared in electrode buffer (25 mM Tris, pH 8.3, 200 mM glycine, and 0.1% SDS) on top of the polyacrylamide gel (14%), and run in the second direction [280]. Gels stained by Coomassie Blue R350 were scanned by ImageScanner III and analyzed using Image Master 2D Platinum 7.0. For the sectional 2DE analysis, this gel was cut into 96 sections with determined coordinates. Each section (~0.7 cm 2 ) was shredded and treated with trypsin. Tryptic peptides were eluted from the gel by extraction solution (5% (v/v) ACN, 5% (v/v) formic acid) and dried in Speed Vac. In the case of a semi-virtual 2DE, the 18-cm Immobiline DryStrip 3-11 NL was cut into 36 equal sections after IEF. For complete reduction, 300 µL of 3 mM DTT and 100 mM ammonium bicarbonate were added to each section and incubated at 50 • C for 15 min. For alkylation, 20 µL of 100 mM IAM were added and samples were incubated in the dark at r.t. for 15 min. The peptides were eluted with 60% acetonitrile and 0.1% TFA and dried in Speed Vac.

ESI LC-MS/MS Analysis
A detailed procedure was described previously [279,280]. Peptides were dissolved in 5% (v/v) formic acid. Tandem mass spectrometry analysis was conducted in duplicate on an Orbitrap Q-Exactive mass spectrometer ("Thermo Fisher Scientific", Waltham, MA, USA). The data were analyzed by Mascot "2.4.1" ("Matrix Sciences", Mount Prospect, IL, USA) or SearchGui [282] using the following parameters: enzyme-trypsin; maximum of missed cleavage sites-2; fixed modifications-carbamidomethylation of cysteine; variable modifications-oxidation of methionine, phosphorylation of serine, threonine, tryptophan, acetylation of lysine; the precursor mass error-10 ppm; the product mass error-0.01 Da. As a protein sequence database, UniProt (October 2014) was used.
Only 100% confident results of protein identification were selected. Two unique peptides per protein were required for all protein identifications. Exponentially modified PAI (emPAI), the exponential form of protein abundance index (PAI) defined as the number of identified peptides divided by the number of theoretically observable tryptic peptides for each protein, was used to estimate protein abundance [283].

Conclusions
For now, proteomics is collecting big data about the human plasma proteome http: //plasmaproteomedatabase.org/index.html accessed on 10 September 2022 [284]. These data include many proteome parameters: their dynamics, different protein presence, abundance, modifications, variations, etc. In the case of cancer, a proteome performs multiple perturbations, where all its components are involved through changes in their levels and modifications. Here, the plasma proteome works as a united entity that executes and reflects the processes in the human body. Accordingly, the profiling of plasma proteomes is a promising and powerful approach to follow these processes. This profiling could combine hundreds of already known plasma biomarkers and has a very promising future in biomedicine as it could disclose information about any abnormal situation in the human body including cancer. There is a big chance that MS-based proteomics will become a part of the routine medical technique [14,285]. In addition to the usual MS analysis of proteins/proteoforms, this technique should include special processing programs allowing conclusions to be made about the human body's state based on these variations in protein/proteoform signatures/profiles (level, PTMs, etc.). In our work, we collected information about the connection of cancers with levels of "classical plasma proteins" and generated their proteoform profiles (Table 1, Supplementary Figure S1). As a next step, similar profiles representing protein perturbations in plasma produced in the case of different cancers should be generated. Moreover, based on this information, different test systems can be developed.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Local Ethics Committee of Petersburg Institute of Nuclear Physics (PNPI) of National Research Center "Kurchatov Institute" (protocol code 02_2020 from 21 April 2020).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: Not applicable.

Acknowledgments:
The work was performed within the framework of the Program for Basic Research in the Russian Federation for a long-term period (2021-2030) (№122030100168-2). Mass spectrometry measurements were performed using the equipment of the "Human Proteome" Core Facilities of the Institute of Biomedical Chemistry (Moscow, Russia).

Conflicts of Interest:
The authors declare no conflict of interest.