Immunoinformatics design of multiepitopes peptide-based universal cancer vaccine using matrix metalloproteinase-9 protein as a target

Abstract

A new approach toward cancer therapy is the use of cancer vaccine, yet the different molecular bases of cancers, reduce the effectiveness of this approach. In this article, we aim to use matrix metalloproteinase-9 protein (MMP9) which is an essential molecule in the survival and metastasis of all types of cancers as a target for universal cancer vaccine design. The reference sequence of MMP9 protein was obtained from NCBI databases. Furthermore, the B-cell and T cell-related peptides were analyzed using the IEDB website and other related soft wares. The best candidate peptides were then visualized using chimera software. Three peptides were found to be good candidates for interactions with B cells (SLPE, RLYT, and PALPR), while 10 peptides were found as good targets for interactions with MHC1 and another 10 peptides founded suitable for interactions with MHC2 with population coverages of 94.77 and 90.67%, respectively. Finally, the immune response simulation and molecular docking were done using the C-IMMSIM simulator and AutoDock Vina to confirm the effectiveness of the proposed vaccine. By the end of this project: twenty-three peptide-based vaccine was designed for use as a universal cancer vaccine which has a high world population coverage for MHC1 (94.77%) and MHC2 (90.67%) related alleles.

Keywords:

• Immunoinformatics
• peptides
• universal cancer vaccine
• IEDB
• matrix metalloproteinase-9

1. Introduction

Cancer is a chronic disease with varying degrees of manifestations characterized by abnormal cell division and high mortality and morbidity rates, in addition, it is considered the second leading cause of death worldwide with an estimated 8 million death annually and 18 million new cases per year [1–5]. Although the cancer survival rate increased dramatically over the years in many countries, yet more work are needed to improve the prognosis of patients suffering from this debilitating disease [6,7]. Cancer types can either be common or rare, depending on the number of cases associated with it. Usually, any cancer with a number of cases below 6 in 100,000 annually is considered a rare type [8]. All types of cancers cause common clinical symptoms including weight loss, fatigue, pain, nausea, vomiting, and constipation, etc. [9]. Additional symptoms are specific to the site of malignancy [10]. Breast cancer is the most common type of cancer in women worldwide [11]. Lung cancer is the second most common cancer in both men and women while colorectal is the third most common type of cancer in both sexes [12].

The classical management for cancer composed of chemotherapy, radiotherapy, and surgery. Recently Immunotherapy started to shows very promising results in the field of cancer management. Different types of immunotherapy are available including cytokine therapy, checkpoint inhibitors, monoclonal antibodies, and therapeutic vaccines [13,14]. With the latter being the center of much research in recent years, cumulating in the licensing of Sipuleucel-T By the U.S. Food and Drug Administration (FDA) as the first dendritic cell-based therapeutic vaccine for prostate cancer in 2010 [15,16]. The vaccine can either be used alone or in combination with other modalities like immune checkpoint inhibitors (ICIs) which shows better results when compared with vaccine monotherapy [17].

Most drug company invests in the common types of cancer like breast and colorectal cancer for financial and cost-effective reasons leaving the rare type of cancer like chondrosarcoma with no effective mode of treatment [18–20]. This calls for a new strategy for the management of cancer as a whole rather than treating different types individually. Furthermore, although many trials of therapeutic and preventive drugs to cancer have been carried out throughout the years, yet none of them succeeded to obtain the same response from all types of cancers, moreover, the FDA has not approved any drug as a universal cancer vaccine agent [21,22]. The reason for this failure is the fact that no effective intervention is available that covers all pathological pathways of the different types of cancer caused by the massive tumor heterogeneity [23].

There are approximately 28 subtypes of matrix metalloproteinase (MMPs). They are responsible for maintaining the components of the extracellular matrix (ECM) and controlling metastasis of cancerous cells. Matrix metalloproteinases-9 (MMP9) gene (MIM 120361) is located at chromosome 20q12-13. It encodes MMP9 protein, a major enzyme for the degradation and absorption of ECM, signaling of wound healing, bone resorption, angiogenesis, and plays a crucial role in various types of cancer progression. Such as breast cancer, colorectal cancer, papillary thyroid cancer, urinary cancers, lymphomas, and gastric cancer [24–27]. In addition, MMP-9 is considered as a biomarker for inflammatory diseases like Rheumatoid arthritis and asthmas [28,29].

MMP-9 is present in normal tissues, however, it is overexpression has been noticed in different malignant tumors like osteosarcoma (73.9%), colorectal carcinoma (69.1%), breast (78.75%), pancreatic cancers (55.17%), and other types of cancers [30–35]. therefore using it as a vaccine target will ensure a universal immune response against most types of cancer. in addition, animal studies in immunized mice with MMP-9 derived peptides resulted in significantly smaller tumor burden and improved immune response against cancer cells with the tested antibodies suppressing migration of both murine and human cancer cell lines in vivo, which suggest a possible outcome in human studies [36].

To the best of our knowledge, this is the first immunoinformatics study to use the human MMP9 protein as an immunogenic target for peptide-based vaccine design for all types of cancer. Peptide-based vaccines are safe, cheaper, and less laboring making them a cost-effective method for vaccination [37,38]. In addition, reverse vaccinology techniques have proven their efficiency in different fields as evidenced by many of the designed vaccines been tested in clinical trials [39–41].

2. Materials and methods

2.1. Sequences retrieval

The whole protein sequence of the MM9 in FASTA format, accessed on November 3, 2019 (NP-004985.2) was retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/protein) [42]. On the other hand, the candidate epitopes were analyzed by IEDB analysis tools (http://www.iedb.org/) [43].

2.2. Identification of conserved regions

Sequences were aligned to identify the conserved regions using the BioEdit tool package version 7.2.5. Furthermore, epitope conservancy investigation for each epitope was predicted by the IEDB analysis resource. (http://tools.immuneepitope.org/tools/conservancy) [44].

2.3 B-cell epitope prediction

B-cell epitope is the part of a vaccine that interacts with the B-cell receptor. The epitopes of interest were investigated using numerous B-cell prediction algorisms to identify the antigenicity, flexibility, hydrophilicity, and surface accessibility, as the selection of the highest immunogenic protein sequence is a prerequisite for epitope-based peptide vaccine design [45].

2.3.1. Prediction of linear B-cell epitopes

BepiPred from the immune epitope database (http://toolsiedb.ofg/bcell/) was used for linear B-cell epitopes prediction [46], with a default threshold value of 0.35. All epitopes with values less than this threshold were chosen.

2.3.2. Prediction of surface accessibility

By using the Emini surface accessibility prediction tool of the immune epitope database (IEDB) (http://tools.immuneepitope.org/tools/bcell/iedb) [47], the surface accessible epitopes were predicted from the available peptides in the conserved region, with a default threshold value of 1.000. We then selected all peptides with values less than this threshold for further analysis.

2.3.3. Prediction of epitopes antigenicity

The Kolaskar and Tongaonkar antigenicity methods were used to determine the antigenic sites in MMP-9 protein. We used a default threshold value of 1.028. It is available at (http://tools.immuneepitope.org/bcell/) [48], all antigenic peptide with values less than this threshold were selected for further analysis.

2.4 T-cell epitope prediction

2.4.1. Mhc class I binding predictions:

Peptide binding to MHC class I molecules was analyzed by the IEDB MHC I prediction tool (http://tools.iedb.org/mhc1.); based on Artificial Neural Network (ANN). This tool has different algorithms to show the ability of the submitted sequence to bind to a specific MHC class 1 molecule [43,49–51]. All epitope lengths were set as nine amino acids, all conserved epitopes that bind to alleles at score equal or less than 100 half-maximal inhibitory concentration (IC50) are selected for further analysis [52].

2.4.2. Mhc class II binding predictions

Peptide binding to MHC class II molecules was analyzed by the IEDB MHC II prediction tool (http://tools.iedb.org/mhcII) [53,54], the three major types of HLA alleles (DR, DQ, and DP) were chosen for analysis based on NN-align, where all peptides with IC50 less than 500 were chosen for a further workout.

2.5. Population coverage

Calculating HLA-alleles distribution among the world population is critical for effective multi-epitope vaccine design; therefore, worldwide population coverage analysis of the above-mentioned MHC classes I and II epitopes was done by using the IEDB tool. (http://tools.iedb.org/tools/population/iedb_input) [55].

2.6. Protein 3D structure visualization

The full 3D structure of MM9 was generated by RaptorX (http://raptorx.uchicago.edu/) [56]. The accession number of the reference sequence is NP-004985.2, then UCSF Chimera version 1.14 was used to visualizing the positions of suggested peptides [57]. The generated protein model was then verified using the Ramachandran plot from the (model panel) tool in chimera software.

2.7. Allergenicity test

AllerTOP v. 2.0 was used to check the Allergenicity of MMP9 protein. This method is based on auto cross-covariance (ACC) transformation of protein sequences into uniform equal-length vectors. It depends on five E descriptors of amino acids (amino acid hydrophobicity, molecular size, helix-forming propensity, the relative abundance of amino acids, and β-strand forming propensity). The proteins are classified depending on training sets containing different known and unknown allergens [58]. It is available at https://www.ddg-pharmfac.net/AllerTOP/index.html.

2.8. Antigenicity test

VaxiJen v. 2.0 was used to check the antigenicity of the chosen peptides with a cut point of 0.5. It is an alignment-free approach for insilico antigen identification, which is based on the ACC method. It has a prediction accuracy ranging from 69 to 97% [59]. The software is available at http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html.

Furthermore, ANTIGENpro from SCRATCH protein predictor server was used to validate these antigenicity results, which is a sequenced based predictive model to predict if certain proteins are protectives [60]. http://scratch.proteomics.ics.uci.edu/.

2.9. Prediction of interferon-gamma inducing epitopes

IFNepitope was used to predict the MHC2 related peptides with a capacity to activate interferon-gamma cytokine productions. It is based on two models: Motif based and SVM based models [61].

The software is available at http://crdd.osdd.net/raghava/ifnepitope/index.php.

2.10. Molecular docking

The MMP-9 protein sequence was visualized in UCSF Chimera. The proposed peptides were then selected and saved as PBD. Furthermore, we choose four peptides from the proposed MHC1 and MHC2 peptides targets, that have the highest population coverages and the largest number of HLA alleles combined to it.

2.10.1. Ligand preparation

The four chosen peptides were prepared using AutoDock Tools 1.5.6 as following, first, we extracted water molecules, then we added polar hydrogen molecules, and lastly, we add charges using the Kollman charge method [62].

2.10.2. Target predation

The 3 D structure of HLA-A*02:01 (6TDR) was retrieved from protein databank (http://www. rcsb.org/ pdb/home/home.do). Afterword, its file was prepared using AutoDock tools 1.5.6.

2.10.3. Insilico docking

Molecular docking was performed using AutoDock Vina 4.0 and PyMOL soft wares, with the default Grid parameters [63]. AutoDock Vina is based on Genetic Algorithm; that combines energy evaluation through grids of affinity potential to calculate the suitable binding position for a ligand on a given protein, It is more accurate when compared with the previous versions of Docking software like AutoDock 4 [64]. The obtained results were retrieved as binding energy. We then used Chimera software to visualize the pose with the least binding energy.

2.11. Immune response simulation

C-IMMSIM model was used to stimulate the immune system response to the proposed vaccine.

We submitted the MMP9 reference sequence to the server and run the simulations with the following prefixed parameters: Random Seed: 12345, Simulation Volume: 10, and Simulation Steps: 100 [65]. http://150.146.2.1/C-IMMSIM/index.php?page=1.

3. Results

3.1. Prediction of B-cell epitope

The metalloproteinase-9 protein (MMP-9) reference sequence was examined by using BepiPred linear epitope prediction, Kolaskar and Tongaonkar antigenicity and Emini surface accessibility methods in IEDB, to calculate the probability of specific regions in the protein which binds to B-cell receptor, being in the surface of the protein and being immunogenic, respectively.

In BepiPred linear epitope prediction, which was the first method used to determine the B-cell epitopes, the average score of MMP-9 peptides was 0.321, with a minimum score of −0.006 and a maximum score of 3.227. Basically, all values that equal to or greater than the default threshold 0.350 were potentially linear epitopes, as shown in Table 1 and Figure 1.

Figure 1. Bepipred linear epitope prediction, the yellow peaks above the red line (threshold) are proposed to be a part of B-cell epitopes and the green peaks are not a part.

Table 1. Describe the predicted linear B-cell epitopes.

No.	Start	End	Peptide	Length
1	21	24	PRQR	4
2	59	67	EMRGESKSL	9
3	79	88	SLPETGELDS	10
4	96	102	TPRCGVP	7
5	129	135	SEDLPRA	7
6	173	184	AEHGDGYPFDGK	12
7	192	205	FPPGPGIQGDAHFD	14
8	221	228	RFGNADGA	8
9	241	269	YSACTTDGRSDGLPWCSTTANYDTDDRFG	29
10	275	288	RLYTQDGNADGKPC	14
11	299	312	YSACTTDGRSDGYR	14
12	318	321	ANYD	4
13	331	343	TRADSTVMGGNSA	13
14	358	369	YSTCTSEGRGDG	12
15	377	390	SNFDSDKKWGFCPD	14
16	425	437	FTEGPPLHKDDVN	13
17	443	515	YGPRPEPEPRPPTTTTPQPTAPPTVCPTGPPTVHPSERPTAGPTGPPSAGPTGPPTAGPSTATTVPLSPVDDA	73
18	544	553	EGRGSRPQGP	10
19	561	565	PALPR	5
20	616	619	ALRS	4
21	641	650	MVDPRSASEV	10
22	657	660	VPLD	4
23	685	690	RSELNQ	6

Twenty-three peptides were identified from MMP-9 protein.

In Emini surface accessibility prediction, the average surface accessibility predictions of the protein were as 1.000, with a maximum of 6.034 and a minimum of 0.031, so all the values equal or greater than the default threshold 1.000 were potentially in the surface. Furthermore in Kolaskar and Tongaonkar antigenicity prediction’s the average of antigenicity was 1.028, with a maximum of 1.307 and a minimum of 0.869; all values equal to or greater than the default threshold 1.028 were potential antigenic determinants. The results of conserved predicted B-cell epitopes that passed both Emini & kolaskar tests are shown in Table 2 and Figures 2 and 3.

Figure 2. Emini surface accessibility, yellow peaks above the red line (threshold) are proposed to be a part B-cell epitopes while green peaks are not a part.

Figure 3. Kolaskar and Tongaonkar antigenicity prediction, yellow peaks above the red line (Threshold) are proposed to be a part of B-cell epitope while green peaks are not a part.

Table 2. List of conserved B-cell epitopes with their surface accessibility and antigenicity scores.

Start	End	Peptide	Length	Conservation	EMINi score	Prediction	KOLSKAR score	Prediction
79	82	SLPE	4	C	1.079	P	1.044	P
275	278	RLYT	4	C	1.331	P	1.048	P
561	565	PALPR	5	C	1.104	P	1.063	P

3.2. Prediction of T-cell epitope

3.2.1. MHC-I binding profile prediction for T cytotoxic cell conserved epitopes

227 epitopes were anticipated to interact with different MHC-1 alleles. The core epitopes (YRYGYTRVA/YGYTRVAEM) were noticed to be the dominant binders (highest population coverage and lowest IC-50) with seven alleles for each (HLA-C*06:02,HLA-C*07:01,HLA-C*12:03,HLA-C*14:02,HLA-B*27:05,HLA-C*07:02,HLA-B*39:01/HLA-C*03:03,HLA-C*14:02,HLA-C*12:03,HLA-C*06:02,HLA-B*08:01,HLA-C*15:02,HLA-C*07:01). Followed by (YLYRYGYTR, WRFDVKAQM, ALWSAVTPL, LLLQKQLSL, LIADKWPAL, KLFGFCPTR, MYPMYRFTE, FLIADKWPA) which binds with [1,3,5,8] alleles (Table 3).

Table 3. List of most promising epitopes that had a good binding affinity with MHC-I alleles in terms of IC50.

Peptide	Start	End	No. of MHC-I allele hits
YRYGYTRVA	50	58	7
YGYTRVAEM	52	60	7
YLYRYGYTR	48	56	8
WRFDVKAQM	633	641	5
ALWSAVTPL	146	154	5
LLLQKQLSL	72	80	3
LIADKWPAL	555	563	3
KLFGFCPTR	324	332	5
MYPMYRFTE	419	427	1
FLIADKWPA	554	562	3

3.2.2. MHC-II binding profile prediction for T helper cell conserved epitopes

479 conserved predicted epitopes were found to interact with MHC-II alleles. The core epitopes (KMLLFSGRRLWRFDV) is thought to be the top binder as it interacts with 18 alleles; (HLA-DRB1*01:01,HLA-DRB1*03:01,HLA-DRB1*04:02,HLA-DRB1*07:01,HLA-DRB1*04:04,HLA-DRB1*08:01,HLA-DRB1*09:01,HLA-DRB1*13:01,HLA-DRB1*11:01,HLA-DRB1*13:02,HLA-DRB1*12:01,HLA-DRB1*15:01,HLA-DRB1*16:02,HLA-DRB3*01:01,HLA-DRB4*01:03,HLA-DRB5*01:01,HLA-DRB3*03:01,HLA-DRB4*01:01). Followed by (GRGKMLLFSGRRLWR, RGKMLLFSGRRLWRF, GKMLLFSGRRLWRFD, TFTRVYSRDADIVIQ, AVIDDAFARAFALWS, FARAFALWSAVTPLT, MLLFSGRRLWRFDVK, GNQLYLFKDGKYWRF, NQLYLFKDGKYWRFS) which binds to [14,15,17,18] alleles (Table 4).

Table 4. List of the most promising core sequence epitopes that had a strong binding affinity with MHC-II in terms of IC50.

Peptide Sequence	Start	End	No. of MHC-II allele hits
KMLLFSGRRLWRFDV	623	637	18
GRGKMLLFSGRRLWR	620	634	17
RGKMLLFSGRRLWRF	621	635	17
GKMLLFSGRRLWRFD	622	636	18
TFTRVYSRDADIVIQ	155	169	14
AVIDDAFARAFALWS	135	149	14
FARAFALWSAVTPLT	141	155	17
MLLFSGRRLWRFDVK	624	638	15
GNQLYLFKDGKYWRF	528	542	14
NQLYLFKDGKYWRFS	529	543	14

3.2.3. Population coverage analysis

Population coverage test was performed, to detect the world Coverage of the epitopes that bind to MHC1 and MHC11 alleles, which founded to exhibit exceptional coverage with percentages of 94.77 and 90.67%, respectively.

Population coverage for MHC1: The interactions between the following 10 peptides with the most frequent MHC class I alleles (YRYGYTRVA, YGYTRVAEM, YLYRYGYTR, WRFDVKAQM, ALWSAVTPL, LLLQKQLSL, LIADKWPAL, KLFGFCPTR, MYPMYRFTE, and FLIADKWPA), represent a considerable coverage against the whole world population. The maximum population coverage percentage over these epitopes worldwide was found to be 62.74% for YRYGYTRVA (Table 5; Figure 4).

Figure 4. Illustrates the global coverage for the top 10 MHC-I peptides (YRYGYTRVA, YGYTRVAEM, YLYRYGYTR, WRFDVKAQM, ALWSAVTPL, LLLQKQLSL, LIADKWPAL, KLFGFCPTR, MYPMYRFTE, and FLIADKWPA). Note: In the graph, the line (-o-) represents the cumulative percentage of population coverage of the epitopes; the bars represent the population coverage for each epitope.

Table 5. Global population coverage of promising epitopes binding to MHC class I alleles.

Peptide	MHC I world coverage (%)
YRYGYTRVA	62.74
YGYTRVAEM	58.08
YLYRYGYTR	54.43
WRFDVKAQM	54.29
ALWSAVTPL	50.41
LLLQKQLSL	46.87
LIADKWPAL	45.42
KLFGFCPTR	43.03
MYPMYRFTE	43.03
FLIADKWPA	42.53

Population coverage for MHC11: In the case of MHC class II, 10 epitopes were found to interact with the most frequent MHC class II alleles (KMLLFSGRRLWRFDV, GRGKMLLFSGRRLWR, RGKMLLFSGRRLWRF, GKMLLFSGRRLWRFD, TFTRVYSRDADIVIQ, AVIDDAFARAFALWS, FARAFALWSAVTPLT, MLLFSGRRLWRFDVK, GNQLYLFKDGKYWRF, NQLYLFKDGKYWRFS), inferring a massive global coverage. The highest population coverage percentage of these epitopes worldwide was that of KMLLFSGRRLWRFDV with a percentage of 83.74% (Table 6; Figure 5).

Figure 5. Illustrates the global proportion for the top 10 MHC-II epitopes (KMLLFSGRRLWRFDV, GRGKMLLFSGRRLWR, RGKMLLFSGRRLWRF, GKMLLFSGRRLWRFD, TFTRVYSRDADIVIQ, AVIDDAFARAFALWS, FARAFALWSAVTPLT, MLLFSGRRLWRFDVK, GNQLYLFKDGKYWRF, NQLYLFKDGKYWRFS). Notes: In the graph, the line (-o-) represents the cumulative percentage of population coverage of the epitopes; the bars represent the population coverage for each epitope.

Table 6. Global population coverage of promising epitopes in isolated MHC class II.

Peptide sequence	MHC II world coverage (%)
KMLLFSGRRLWRFDV	83.74
GRGKMLLFSGRRLWR	83.46
RGKMLLFSGRRLWRF	83.46
GKMLLFSGRRLWRFD	83.14
TFTRVYSRDADIVIQ	81.42
AVIDDAFARAFALWS	80.82
FARAFALWSAVTPLT	79.23
MLLFSGRRLWRFDVK	78.71
GNQLYLFKDGKYWRF	78.50
NQLYLFKDGKYWRFS	78.50

3.2.4. Antigenicity analysis

Antigenicity analysis of the 10 chosen peptides of MHC1 and MHC2 by VaxiJen software shows four peptides to be probably antigenic in MHC1 (YLYRYGYTR, WRFDVKAQM, LIADKWPAL, and MYPMYRFTE) while only one peptide was found to be antigenic in MHC2 (RRLWRFDVKAQMVDP) (Table 7). Furthermore, ANTIGEN pro predicts MMP-9 protein probability of antigenicity to be 0.827661.

Table 7. List of antigenic peptides inMHC1 and MHC2 as predicted by VaxiJen software.

MHC	Peptide Sequence	Prediction	Score
MHC1	YLYRYGYTR	Probable antigen	0.7645
	WRFDVKAQM	Probable antigen	0.7671
	LIADKWPAL	Probable antigen	1.2157
	MYPMYRFTE	Probable antigen	0.7337
MHC2	RRLWRFDVKAQMVDP	Probable antigen	(0.5585)

3.2.5. Prediction of interferon-gamma inducing epitopes:

IFNepitope server showed three peptides in MHC2 with an ability to induce interferon-gamma production (GPFLIADKWPALPRK, GRGKMLLFSGRRLWR, and FARAFALWSAVTPLT) (Table 8).

Table 8. List of MHC2 related peptides that can induce interferon-gamma production.

Peptide	Prediction	Score
GPFLIADKWPALPRK	Positive	1
GRGKMLLFSGRRLWR	Positive	0.047075488
FARAFALWSAVTPLT	Positive	0.40888389

3.2.6. Structure modeling and validation by chimera

The MMP-9 protein structure was predicted by RaptorX and then analyzed by chimera software, moreover, the chosen peptides related to B cell, MHC1, and MHC2 peptides were then visualized (Figures 6–8) Furthermore, the modeling was verified using the Ramachandran plot analysis tool in chimera software, which shows that most of the amino acid residues are located in the allowed region (Figure 9).

Figure 6. Present 3D structure of metalloproteinase-9 protein visualizing the top 10 T-cell peptides binding to MHC class I using chimera (version 1.14), yellow color indicate the chosen peptides while the purple color indicates the rest of the molecules.

Figure 7. Present 3D structure of metalloproteinase-9 protein visualizing top 10 T-cell peptides binding to MHC class II, using chimera (version 1.14), yellow color indicate the chosen peptides while the purple color indicates the rest of the molecules.

Figure 8. Present 3D structure of metalloproteinase-9 protein visualizing the most promising B-cell peptides that have high surface accessibility and antigenicity scores, using chimera (version 1.14), yellow color indicate the chosen peptides while the purple color indicates the rest of the molecules.

Figure 9. Ramachandran plot analysis of MMP9 protein showing most of the torsion angles located at allowed region (the blue dots represent torsion angles; the green lines indicate the allowed region. (Phi) ϕ And (psi) ψ are torsion angles. The torsion angle about the N–C bond is called ϕ and that about the C–C bond is ψ. This analysis predicted by chimera version 1.14.

3.2.7. Conservation analysis

BioEdit analysis showed most of the amino acids in MMP-9 proteins to be conserved, except in three positions (110, 279, and 574) where the original wild amino acids in these position changes to Leucine, Arginine, and Proline, respectively (Figure 10).

Figure 10. MMP9 sequences conservation test by BioEdit software showing three non-conserved regions at positions 110, 279, and 574. The continuous colored dots represent conservation.

3.2.8. Characterization of the immune response of the vaccine

To investigate the immune responses created by the final chimeric vaccine construct, the immune simulator C-IMMSIM Online server-generated such simulations, which resemble the real responses produced by the immune system. These responses suggest a high spike in the initiation of secondary immune responses. Relatively, high levels of IgM were recorded in the primary response in the first days of vaccination, preceded by the surge in B-cell populations. Furthermore, the IgG1 + IgG2, IgG + IgM antibodies along with IgM were found to be increased along with the formation of T and B memory cells (Figure 11(A–C)). As the memory further intensifies, a projected response was reported in the cytotoxic and helper cell populations (Figure 11(D,F,G)). Furthermore, cytokines like IL-6, IL-12, and interferon-g are reported to increase in the first days after the introduction of the vaccine to reach a level of 250,000 ng/ml by the end of the second week. Moreover, IL-2 was noticed to peak after 7 days of vaccine administration (Figure 12).

Figure 11. Immune response to MMP-9 as predicted by the C-IMMSIM Online server. The plot shows the simulated levels of memory cells, B cell, and T cells after the introduction of MMP-9 protein to the human body. Act = active, Intern = internalized the Ag, Pres II = presenting on MHC II, Dup = in the mitotic cycle, Anergic = anergic, Resting = not active.

Figure 12. Concentration of cytokines and interleukins. The inset plot shows the danger signal together with leukocyte growth factor IL-2.

3.2.9. Docking

Four peptides were chosen for docking with HLA-A2A01 the result was: nine possible positions for each peptide with MHC1 and MHC2 proteins, we then selected the positions with least energy as our final model. The results of the least energy exposure for MHC1 peptides were as following: ALWSAVTPL: −9.7, LLLQKQLSL: −8.6. Furthermore, the results of the least energy exposure for MHC2 peptides were as following: KMLLFSGRRLWRFDV: −8.8, GRGKMLLFSGRRLWR: −7.8. The following are 3D representations of the docking of four peptides with MHC1 and MHC2 proteins (Figures 13–16).

Figure 13. First peptide ALWSAVTPL docking with HLA-A*0201, showing the position with the least energy exposure (−9.7). Blue and cyan colors indicate MHC protein, while white and red color represents the binding peptide. The docking was done using AutoDock Vina and visualized using Chimera version.1.14.

Figure 14. Second peptide LLLQKQLSL docking with HLA-A*0201, showing the position with the least energy exposure (−8.6). Blue and cyan colors indicate MHC protein, while white and red color represents the binding peptide. The docking was done using AutoDock Vina and visualized using Chimera version.1.14.

Figure 15. Third peptide KMLLFSGRRLWRFDV docking with HLA-A*0201, showing the position with the least energy exposure (−8.8). Blue color indicates MHC protein, while white and red color represents the binding peptide. The docking was done using AutoDock Vina and visualized using Chimera version 1.14.

Figure 16. Fourth peptide GRGKMLLFSGRRLWR docking with HLA-A*0201, showing the position with the least energy exposure (−7.8). Blue color indicates MHC protein, while white and red color represents the binding peptide. The docking was done using AutoDock Vina and visualized using Chimera version.1.14.

4. Discussion

Therapeutic and preventive multi Epitope peptides vaccines are the next generation of cancer immunotherapy. Production and testing of vaccines are expensive and time-consuming procedures, hence the development of immunoinformatic studies to tackle these challenges associated with the classical production of vaccines [66,67]. Many researchers have tried to design universal cancer vaccines using different approaches, yet until now, the FDA has not approved any of these drugs [68]. The current work used immunogenic protein (matrix metalloproteinase-9) that is essential for the survival of all type of cancer as a target for vaccine design, where three Peptides found to be a good candidate for interactions with B cells, in addition, 10 peptides found as a good target for interactions both MHC1 and MHC2. The different soft wares used for the analysis are summarized in Figure 17.

Figure 17. Workflow of the paper: showing the different soft wares used.

Most cancer vaccines depend on T cell interactions but recent research indicates the significant role B cell plays in the immunology of cancer, B cell produces antibodies and help in the regulation of other immune reaction [69]. Furthermore, several B cell-based cancer vaccine studies have shown Promising results in chronic lymphocytic leukemia and other types of tumors [70]. In this article, Linear B cell epitopes of matrix metalloproteinase-9 were analyzed by the BepiPred linear epitope prediction tool. After checking for conservation using BioEdit software (Figure 10). These epitopes were further checked for surface accessibility and antigenicity by using the Emini surface accessibility prediction method and Kolaskar and Tongaonkar antigenicity method respectively (Figures 1–3) The following peptides have the highest score and were approved as the most promising B-cell epitopes (SLPE, RLYT, and PALPR) (Tables 1 and 2).

T-cell receptor can only identify epitopes if it is presented on major histocompatibility complex (MHC), Epitopes bind to MHC through the binding groove of MHC class I and class II by interacting with the amino acids residues located in the pockets on the floor of MHC [71,72].

Cytotoxic T cell (CD8) can only recognize antigen when it is presented on MHC1, and so we analyzed the peptides that have the highest population coverage and the best binding MHC1 and found the following 10 candidates (YRYGYTRVA, YGYTRVAEM, YLYRYGYTR, WRFDVKAQM, ALWSAVTPL, LLLQKQLSL, LIADKWPAL, KLFGFCPTR, MYPMYRFTE, FLIADKWPA) (Table 3; Figure 4). It is interesting to note that among the associated alleles related to these peptides, we found the most frequently reported allele in Caucasians which is HLA-A*0201 [73]. This further proves the significant coverage of these peptides.

In the case of T-helper cells, it can only recognize antigen when it is presented in MHC2. Therefore, we predicted the peptides of MMP9 that can firmly bind with MHC2, which also hold the highest population coverage and end up with the following 10 peptides sequence (KMLLFSGRRLWRFDV, GRGKMLLFSGRRLWR, RGKMLLFSGRRLWRF, GKMLLFSGRRLWRFD, TFTRVYSRDADIVIQ, AVIDDAFARAFALWS, FARAFALWSAVTPLT, MLLFSGRRLWRFDVK, GNQLYLFKDGKYWRF, and NQLYLFKDGKYWRFS) (Table 4; Figure 5).

The population coverage of the epitopes and their respective HLA alleles are crucial, as HLA alleles are highly polymorphic in nature [74]. We aim for this vaccine to benefit patients from different parts of the world, therefore, we analyzed the whole word population coverage for our predicted peptides, where we found MHC1 and MHC2 related peptides to have population coverage of 94.77 and 90.67%, respectively (Tables 5 and 6). Therefore, our vaccine is assumed to elicit a strong comprehensive immune response against most types of cancer in the majority of the vaccinated population.

All peptides were visualized using Chimera software (Figures 6–8) which shows the predicted structures and positions of these peptides in MMP9 protein (yellow color), then Ramachandran plot analysis shows most of the amino acids residue located in the allowed regions which validate the accuracy of the generated structure (Figure 9) [75].

Allergenicity is an important factor in vaccine design as the allergic reaction could be fatal to the patients and hence we tested the allergenicity of our proposed vaccine using AllerTOP software which confirms the nonallergenic of the vaccine [76]. Furthermore, the antigenicity is a significant factor to put into account when choosing a target for vaccine design [77], therefore we tested the antigenicity of mhc1 and mhc2 peptides using VaxiJen software which shows 4 peptides to be antigenic in mhc1 (YLYRYGYTR, WRFDVKAQM, LIADKWPAL, MYPMYRFTE) and one peptide to be antigenic in mhc2 (RRLWRFDVKAQMVD) (Table 7). In addition, the antigenicity of the whole protein was checked using ANTIGENpro which is more accurate and specific than VaxiJen and the resulted probability (0.827661) indicated the high antigenicity of the protein [60,78].

Interferon gamma is a cytokine that plays an essential role in innate and adaptive immunity. It is produced from T cell upon activation by specific peptides which lead to an improvement in cancer elimination by inducing ischemia and improving recognition of the tumor cells by CD4 and CD8 cells [79]. we tested the ability of our chosen peptide to induce secretion of interferon-gamma through IFNepitope software which indicates that there are 3 peptides in MHC 2 to positively affect the interferon-gamma production (GPFLIADKWPALPRK, GRGKMLLFSGRRLWR, FARAFALWSAVTPLT) (Table 8).

Molecular docking and validation were done using AutoDock Vina after protein and ligand preparation using AutoDock tools. The docking shows the best combination possible between the peptides and the MHC protein. Here we choose the best four peptides for docking (ALWSAVTPL: −9.7, LLLQKQLSL: −8.6, KMLLFSGRRLWRFDV: −8.8, and GRGKMLLFSGRRLWR: −7.8.) and we showed the best model with least energy exposure using Chimera. This will help in future studies regarding the design of the MMP-9 vaccine (Figures 13–16).

Immune response simulation was done using the C-IMMSIM server which is an agent-based simulator of the immune response against viral and drugs. After submitting the reference sequence, we have the result of both cellular and humoral response, which shows an increase in IgM antibodies after less than 5 days since the beginning of vaccination followed by fast drop then a rise in IgG antibodies after it, which is a known immune response phenomenon in vaccination [80] (Figure 11(A–G)).

Interesting findings were the rise in memory cells for more than a month since the beginning of infection suggesting a possibility of developing a lasting immunity after using this vaccine (Figure 11(A,D,F)).

The associated cytokine changes prove the vast immunogenic reactions induced by the vaccine especially the rise in IL2, in 7 days after the vaccine induction, which plays a significant role in enhancing the immune response against the cancer antigens through potentiating the role of natural killer cells NKL, B cell, and T cell, therefore inducing regression in cancer patients (Figure 12) [81,82].

In a study done by Roomi et al. where they used MMP9 as an anticancer vaccine in mice, they found an 80% reduction in tumor volume after using this vaccine [83]. Which further proves the predicted effectiveness of our proposed vaccine. We took further steps in our work to analyze this target using the human database.

Previous research by Dosset et al. designed a universal peptide-based cancer vaccine using telomerase reverse transcriptase (TERT) which showed improve response only against CD4 and CD8. With no testing for the possible response from B cells which was covered in our current research [84].

This study is limited due to the few numbers of reference sequence we were able to acquire from NCBI, along with the six alleles (HL-DRB3*03:01, HL-DRB3*02:01, HL-DRB4*01:03, HL-DRB3*01:01, HL-DRB4*01:01, HL-DRB5*01:01) that was not applicable for prediction in IEDB. In addition to the limited reported therapeutic activity of peptide vaccines in general [85].

In the end, 23 peptides were chosen as a target for universal cancer vaccine design, yet more clinical studies are required to confirm and validate these results, in order to move another step towered the goal of eradicating cancer.

5. Conclusion

The current work designs multi epitopes peptide-based universal cancer vaccine using MMP9 protein as a target where three Peptides were found to be good candidates for interactions with B cells, and 10 peptides were found as a good target for interactions with each MHC1 and MHC2.

Acknowledgment

The authors acknowledge the Deanship of Scientific Research at the University of Bahri for supportive cooperation.

Disclosure statement

The authors declare that no conflict of interest exists in this work.