Immunogenetics, as the meeting point of two exciting fields of immunology and genetics, is a new but rapidly expanding field of science studying this immune polymorphism in order to understand the governance of genetics on the immune system [14, 27, 28].

Although the term “immunogenetics” was used earlier [29], the first milestone in the history of immunogenetics was coincident with the failed study of blood transfusion in 1952 [30]. This failure resulted in the discovery of HLA system [14, 31], which attracted the attention of biomedical researchers to interindividual differences in the immune system. From that point on, for decades, investigators tried to associate different complex diseases with various HLA types using serological methods [32, 33]. However, modern immunogenetics required more than one century of biomedical advances remarked by Mendel’s laws of heredity in 1865 [16, 34], discovery of chromosomes as the cellular basis of heredity in 1902, discovery of DNA double helix as the molecular basis of heredity in 1953 [35], decoding the genetic codes, and last but not least the completion of Human Genome Project in April 2003 [16, 36, 37]. Human Genome Project not only contributed to the discovery of genetic polymorphisms but also provided a infrastructure for other large-scale projects like International HapMap Project and “1,000 Genomes Project” [38]. Discovery of approximately 25–35 % of estimated nine to ten million SNPs is just one of the uncountable achievements of such projects [14, 37–39]. Genetic polymorphisms in the immune system contribute to a large part of the interindividual variation in immune response and today, immunogenetic studies have provided a vast knowledge of the effects of immune polymorphism on the host defense. However, just the estimation that there is one SNP per every 290 bp shows that there is much more to be brought to light [38, 39].

Along with the concert of conceptual advancements, tools employed in this field have changed in order to gather immunogenetic information more accurately, in less time and less cost [14]. Twin studies recruit twins in order to remark the importance of genetic component in susceptibility to traits and diseases [16, 40]. The result of such studies provides a rough estimation of genetic contribution to interindividual differences in immune system by comparison of concordance rates of immune traits between monozygotic and dizygotic twins [16, 33, 40]. The higher the concordance difference is, the greater the heritability [7, 16].

Upon introduction of immune polymorphism, several association studies tried to show the contribution of specific genes using the candidate gene approach or hypothesis-driven approach [16, 41]. This approach includes looking into the differences between patients and controls in allele frequencies of SNPs in genes selected based on the known pathophysiologic pathways of the disease. These studies at first employed restriction enzymes to identify specific SNPs called restriction fragment length polymorphisms (RFLPs) in the restriction site of the enzyme [42]. This approach is also known as a reductionist approach, since studies employing this approach investigate only a few genes and polymorphisms at a time [16, 41, 43].

In the early 1990s, discovery of hundreds of informative microsatellites provided the possibility of a dramatic change in the approach of immunogenetic studies from a hypothesis-driven approach to positional approach [4, 16, 44]. In this approach, studies known as genome-wide association studies (GWAS) mainly aim to identify the genome regions bearing disease-associated genes and to localize causal genetic variants of disease as accurately as possible [44, 45]. Therefore, in this approach, new hypothesis are generated after making thousands of unbiased observations [4, 33, 41, 44]. They are especially helpful in order to find unexpected genes as representatives of unknown disease-related pathway [4, 16]. In mid-1990s, early GWASs employed informative microsatellite markers distributed evenly in the 23 chromosomes and investigated their aggregation in multi-case families and large pedigrees identified major susceptibility loci for complex diseases [4, 42, 44].

By introduction of linkage disequilibrium (LD) defined as the coinheritance of alleles of a block of neighboring SNPs; in 2002, the International HapMap Project, as a global movement, began to identify these blocks (known as haplotypes) and pattern of LD in the human genome [38, 39]. LD results in organization of genetic variation in haplotype blocks with strong LD separated by recombination hotspots [16, 39]. The information from this project provided the immunogenetic scientists with the most suitable SNPs for genotyping in order to indirectly gather as much as information about the genome variation of an individual [16, 46]. These SNPs, which are representative of a block of SNPs, are known as tagSNPs. The extent of LD in a region determines the number of tagSNPs required to cover a region. The lower the LD is in a region, the higher number of tagSNPs are needed and therefore the higher the cost of genotyping the region is [47]. Nowadays, availability of high-throughput gene technologies such as gene chips or microarrays has enabled investigators to genotype cost-effectively, rapidly, and almost effortless hundreds of thousands to millions of SNPs at the same time [4, 33, 41, 44]; therefore this approach is also known as “nonreductionist” approach [4]. These technological advancements were employed in community-based and large-scale GWASs in order to identify trait-associated regions with higher resolution. The results of such studies is a trait-associated SNP (TAS) as a representative of the true casual variant which might be each of the known and unknown variants in whole TAS block. The TAS block is defined as all known and unknown polymorphisms in strong LD with the tagSNP [4, 16, 48]. Therefore, LD along with technological advances turned SNPs, the most common and more importantly the most stable genetic variations in human DNA, into application [49].

However, there are major limitations in GWASs to be overcome.

After identification of associated TAS blocks by GWASs, the actual functional variant in the associated TAS block can be found by further genetic association studies employing more accurate low-throughput technologies and other SNP markers in order to finely map the associated genes and alleles in the associated TAS block [44]. In these studies, allele frequencies of polymorphisms are compared in groups of cases and controls. However, results of such association studies are often contradictory due to the heterogeneous nature of the cancers, numerous gene–gene and gene–environment interactions [58, 59]. In addition, another source of discrepancy between these studies is the limitation in study design. For example, using hospital-based controls can result in a serious selection bias since polymorphisms under investigation might have association with the diseases that hospital-based controls may have [60, 61]. Moreover, some association studies failed to consider other genetic and environmental risk factors such as socioeconomic status, nutritional statues, smoking patterns, etc. [60]. Lacking such information may cause serious confounding bias [62]. Therefore, in order to get the most benefit from results of genetic association studies and to systematize their findings, employing meta-analyses as a powerful statistical method is essential [26, 63]. Meta-analysis by pooling the results of old studies allows us to see the whole picture of the effect of a certain polymorphism [26].

Regardless of interspecies differences, there are similarities in cancer development between humans and rodents, and therefore mouse studies are a complementary tool for genetic association studies within human population [6, 64, 65]. Numerous genetically engineered mouse (GEM) models provide a simplified model of various cancers with controllable genetic and environmental background in which the effects of a unique polymorphism on the malignancy can be studied [6, 66].

Exact mechanism of action of polymorphisms can be identified using different bioinformatic tools and in vitro studies [24]. Numerous bioinformatic online and offline tools are available which can predict the effect of polymorphisms by considering amino acid biophysical properties, active site residues, metal and lipid binding sites of gene product, TFBSs, splice sites and its regulatory motifs, miRNA binding sites, and PTM sites (Table 17.1) [24]. However, bioinformatics is limited by the extent of our knowledge [22, 24].

Different in vitro methods are developed to identify functional polymorphisms. The most important ones are reporter gene assay and electrophoretic mobility shift assay (EMSA) (Figs. 17.4 and 17.5) [22]. The reporter gene assay employs a reporter gene with a quantifiable product and clones the promoter of interest in its upstream [22, 67, 68]. Therefore, quantification of reporter gene product can provide information about the promoter strength [22, 67, 68]. On the other hand, EMSA can measure the effect of different polymorphisms on the affinity of TFBS sequence for different transcription factors. In these studies, double-stranded oligonucleotide containing the polymorphism of interest is mixed with nuclear extract with various transcription factors [22, 69, 70]. Higher affinity for these factors results in the formation of more protein–DNA complex resulting in retardation of mobility in electrophoresis [22, 69, 70].

The results from immunogenetic studies should always be interpreted with consideration of information from immunogenomics and immunoproteomics [33]. It should be noted that information from each type of study i.e., GWASs, genetic association studies, in vitro and mouse studies and bioinformatics, are just pieces of the complex puzzle of immunogenetics and cancer. No individual method is precise enough to see the final picture (Fig. 17.6).

Human leukocyte antigens are specialized elements of the immune system in recognition of self from non-self. HLA is responsible for presenting Ags to T cells and therefore serves as a door to the specific immune system. HLA class 1 Ags are on the surface of almost all nucleated cells and generally present processed endogenous antigens to CD8⁺ cells [13, 76]. Presentation of abnormal Ags derived from intracellular pathogens or malignant transformations potentially initiate a cytotoxic T lymphocyte (CTL) response and consequently target cell lysis [77]. By their interaction with killer cell immunoglobulin-like receptors (KIRs) on the surface of natural killer (NK) cells, HLA class 1 antigens regulate lytic activity of NK cells. Therefore, any change in either in expression or structure of HLA class 1 profoundly influence T and NK cell-mediated immunity [10].

On the other hand, HLA class 2 Ags are exclusively expressed on the surface of professional antigen-presenting cells (APC) and present processed exogenous Ags to T helper (Th) cells. Following presentation of unfamiliar Ags and in the presence of appropriate costimulatory molecules, Th cells activate effector elements of the immune system [13, 77].

Both classes of Ags comprise an intracellular, transmembrane, and an extracellular part which includes highly polymorphic antigen binding groove. From the evolutionary view, this high variety favors the chance of heterozygosity and consequently Ag presenting potential for each individual along with a significant increase in the general repertoire of the whole specie for Ag presentation [14, 77].

HLA loci, located in 6p21.3 region, occupy only a small part of major histocompatibility complex (MHC) genetic system which is home to at least 220 genes [78, 79] (Fig. 17.7). MHC is divided into three classes of genes distributed from centromere to telomere. Class 2 with 0.9 mb is the nearest one to the centromere; class I with 1·9 Mb is near telomere, and class 3 with 0·7 Mb lies in between [80]. The first two classes encode for HLA class 1 and 2 and the third class consists of a group of genes encoding some members of the complement system, some cytokines like tumor necrosis factor alpha (TNF-α), heat shock proteins (HSP) and an enzyme called 21-OH hydroxylase [31, 80].

In class 1, there are three highly polymorphic classic genes known as HLA-A, HLA-B and HLA-C, while there are numbers of nonclassical genes known as HLA-E, HLA-F and HLA-G [81, 82]. Class 1 genes encode the highly polymorphic heavy chain of HLA class 1 (45 kDa) which later joins the non-polymorphic B2 microglobulin encoded by chromosome 15 [81, 82]. Classic genes consist of eight exons, but the most important exons are exons 2 and 3 encoding for peptide binding groove. Other exons encode for transmembrane region and cytoplasmic tail [31, 83]. Beside these highly polymorphic classic HLA class 1 genes, there are three other HLA genes in class 1 known as HLA-E, HLA-F and HLA-G which are more conserved. Most probably, they are not involved in Ag presentation but in interaction with more conserved parts of the immune system. For example, HLA-E, which is minimally polymorphic, regulates cytotoxic activity of NK cells by interacting with CD94/NKG2 lectin-like receptors. The conservation within this gene guarantees that there is a constant protection for healthy cells in most people and provides a minimum safeguard for autoimmunity [32, 84, 85]. Some of them like HLA-G are expressed on trophoblastic cells and placental chorionic endothelium and induce immune tolerance during pregnancy [81, 86–90].

Class 2 consists of classic genes called DP, DQ and DR and nonclassic genes known as DM and DO. Classic genes encode for one highly polymorphic beta chain (26–28 kDa) and a less polymorphic alpha chain (33–35 kDa) [80]. Therefore, there are six classic D genes in this region. Genes for alpha chain consist of five exons, while beta chains are encoded by six exons. The exons 2 and 3 in both set of genes are responsible for encoding peptide binding domains [31].

HLA class 1 and 2 genes are the most polymorphic genes in the human genome with 2,365, 3,005 and 1,848 alleles for HLA-A, HLA-B and HLA-C, respectively, and 2,156 alleles for class 2 genes (based on IMGT/HLA database, release 3.13 on July 2013) [91]. This high polymorphism is mostly clustered in several hypervariable blocks in exons 2 and 3 which are responsible for encoding antigen binding groove. Therefore, a unique combination of sequence motifs in these hypervariable regions determine each allele [13]. This genetic structure is accompanied by high LD not only between HLA genes but also non-HLA genes constituting extended haplotypes [92]. The majority of polymorphisms in hypervariable regions result in amino acid substitutions in peptidebinding grooves, which in turn dramatically changes Ag binding affinity of the final product [13]; on the other hand, variants in noncoding regions, influence transcription, translation, and splicing and thereby expression levels [77].

Nowadays, with a few exceptions, HLA alleles are named by six or even eight digits. The first two digits are representative of the serological family the allele belongs to, while the third and fourth digits distinguish between different sequences affecting amino acid sequences. The next two digits are identifiers of synonymous polymorphisms, and seventh and eighth digits are used to distinguish intronic polymorphisms or ones located into untranslated regions [93].

HLAs are involved in cancer immunity and therefore in susceptibility and prognosis mainly by presenting certain Ags known as tumor-associated antigens (TAA). TAA are the first contact of malignant cells with adaptive immunity. Since introduction of the first TAA in melanoma patients in 1991, a broad heterogeneous group of Ags were discovered and associated with different malignancies. This heterogeneous group can be divided in to four classes of Ags [8, 94]:

Nowadays, hundreds of HLA association studies prove that HLA alleles are important elements in predisposition to cancer. Seven mechanisms are suggested for complex relationship of HLA genotypes and susceptibility, prognosis, recurrence, and clinical response to immunotherapy and tumor vaccines:

HLA has a history as long as immmunogenetics itself. An observation of transfusion failures in 1952 paved the road to the discovery of the first HLA allele by Jean Dausset in 1958 [111]. Since 1958, there was a continuous international effort in order to share experimental data and HLA typing technologies, identify new HLA alleles and serotypes, and uncover the role of HLA system in pathogenesis of numerous diseases [31]. The result of such effort was the identification of over 9,500 alleles for HLA class 1 and 2 over a short period of four decades [31]. Along with the discovery of new alleles, the first nomenclature committee was held in 1987 followed by several nomenclature committees to unify the nomenclature and classification [31].

Early studies employed low-resolution serological methods which detected HLA on T cells or B cells [112]. Although these serological methods were subject to huge development in detection methods from complement-dependent cytotoxicity test to ELISA method, flow cytometry, and Luminex technique, the real breakthrough in HLA association studies was the introduction of PCR and high resolution DNA-based typing methods [31]. This technology allowed not only detection of high HLA polymorphisms with higher sensitivity and specificity but also the detection of new alleles with more flexibility by simply adding new probes to the old panels [113]. Nowadays, the old DNA-based method employing PCR-RFLP has been replaced by more rapid tests [113]. Generally, they either identify PCR products containing hypervariable regions by hybridization with sequence-specific probes (SSO) or employ sequence-specific primers (SSP) to identify variants as part of PCR process itself [13, 31, 114]. The latter was extensively used back in mid-1990 [13, 31, 114]. Even though aberrant typing as a sign of new allele can be followed by direct DNA sequencing, both methods are ineffective in case there is a new allele [13]. Later this limitation was overcome by polymerase chain reaction-sequence-based typing which can directly detect the sequence of alleles. In this method which is based on dye terminator chemistry, dye bounded 2,3 dideoxynucleotides are used as substrates for PCR process. Randomly addition of labeled dideoxynucleotides, and consequently, a stop in elongation of DNA chain result in the development of numerous DNA fragments with different sizes. These DNA fragments can easily be separated by capillary electrophoresis, and the ending dideoxynucleotides can be identified by specific fluorescence emitted from the related dye.

In parallel, huge efforts were made to understand the role of these alleles in etiology and natural history of several diseases. In oncology, the first association was found in HL in 1967 [32]. This finding triggered a series of HLA association studies on different cancers worldwide. The fruit of this global movement was finding association between HLA alleles and susceptibility to several hematological malignancy including HL, NHL, childhood acute lymphoblastic leukemia, Kaposi’s sarcoma, chronic myeloid leukemia (CML), and also non-hematological malignancies including nasopharyngeal carcinoma, thyroid cancer, renal cell carcinoma (RCC), cervical cancer, and both melanoma and non-melanoma skin cancers [13, 115]. Moreover, investigations on natural history of cancers showed relationship of several alleles from both classes with mortality in ovarian cancer, non-small cell lung carcinoma, head and neck squamous carcinoma, and local recurrence in melanoma [73, 96, 100]. Several studies showed importance of HLA context in the outcome of immunotherapy and tumor vaccines in melanoma, RCC, cervical carcinoma and CML [73, 95, 110, 116].

Although the result of such studies was inconsistent in some cases, most studies pointed to the undeniable role of HLA polymorphism in susceptibility, prognosis, natural history, and response to immunotherapy in different cancers [32]. These past experiences emphasize that a prestigious HLA association study is a complex art rather than a simple case-control study and several factors should be considered in interpreting their results. In this regard, results of meta-analysis of these association studies are more reliable (Table 17.2).

Indeed, immunogenetic studies are deeply influenced by technological advances. Low-resolution serologic HLA typing was one of the major limitations in early studies [83]. Serologic typing is only enabled to identify the family of alleles. This family often comprises a heterogeneous group of alleles with different affinities and different potential for Ag presentation. Since distribution of alleles belonging to the same serotype is different in various populations, such studies often obtained conflicting results in different populations. One of the best historical examples is HLA association studies in nasopharyngeal carcinoma (NPC).

NPC, as an epithelial carcinoma of the head and neck origin, was one of the main focuses of early HLA association studies. Early serological studies showed an association between HLA-A2 and NPC-in Chinese population, while studies in Caucasians found HLA-A2 as a protective allele for both NPC and EBV-associated HLA [106, 120–124]. Later, higher-resolution studies showed HLA-A*02:07, a common allele in Chinese population but rare among Caucasians, as the main risk factor, while HLA-A*02:01, a common allele in Caucasians, was shown to be the actual protective factor in this population [125, 126]. It is possible that future studies employing higher-resolution methods reveal even new causal variants within the current associations.

Various environmental and genetic factors play roles behind scenario of cancer, and malignant transformation is the result of a complex interaction between these factors. It is often the case that certain genetic factors need certain environmental factors to play their role in pathogenesis of cancer. The role of environmental factors in HLA association studies is more prominent in virus-associated malignancies like HL, NPC and cervical cancer. Each virus has different strains with different Ags and the prevalence of these strains is not the same in different populations. Each strain is best presented by certain HLA alleles. Therefore, one HLA allele efficient for presenting Ags of one population’s prevalent strain may not present Ags of another population’s prevalent strain efficiently [83]. Such a phenomenon might be extended to other environmental factors like virus prevalence, viral load, diet, cigarette smoking, and socioeconomic status, all of which are highly dependent on the population under study [74, 127]. For instance, pathogenesis of cervical cancer is dependent on persistent infection with high-risk human papillomavirus (HPV) and this risk factor itself is highly related to socioeconomic status, sexual relationship, and prevalence of high risk variants in the region [127, 128].

MHC region is home to more than 200 genes beside classic HLA genes. Due to the low recombination rates, these genes are often in strong linkage disequilibrium together [78]. This strong LD can complicate finding the actual causal allele. The problem gets worse when the causal allele is an unknown allele in strong LD with the associated allele. This limitation can be overcome by whole genome sequencing (WGS) of the region in close proximity of the associated allele [101]. One example is the association of NPC with HLA-A*0207 and HLA-B*4601 which are in strong LD. In this case, either allele, both of them, or even a third allele in LD with both of them might influence the pathogenesis of NPC [126].

Some studies reported extraordinary LD in MHC region between alleles from one class and alleles of other classes and even non-HLA genes. This extraordinary haplotypes are known as extended haplotypes [83]. Thus, in interpreting results of HLA association studies or design of one, non-HLA genes such as the transporter associated with Ag processing (TAP) MHC class I chain-related A (MIC-A), heat shock proteins (HSP), and TNF-α which are located nearby or within the classic HLA genes should be considered [78, 83]. These extended haplotypes are especially of importance in immunogenetic studies of cancers, since numerous elements of the immune system are in the front line of defense against cancer.

For instance, the ancestral haplotype 8.1 (AH 8.1: HLA-A*01-B*08-Cw*07-DRB1*03-TNF-G308A), in which HLA alleles are in LD with TNF-α, is the most frequent extended MHC haplotype in Caucasian populations [109]. Primarily, this extended haplotype was associated with clinical course of NHL [75, 109]; however, later studies showed that polymorphism in TNF-α gene has a more prominent effect in this association compared to Cw*07 and DRB1*03 alleles [8, 75]. In this case, polymorphisms in TNF-α promoter influence TNF-α expression levels. TNF-α level consequently affects the extent of immune activation upon tumor challenge. In addition, increased TNF-α impairs Ag presentation potential of APCs and by its effect on cytokine profile results in a bias toward Th2 immune responses [75]. All these factors can contribute to the exacerbation of systemic symptoms, anemia, hypoalbuminemia, and poor outcome [8].

Another example is the association of HLA-A*03 and chronic myeloid leukemia (CML) [78]. A translocation between t(9;22)(q34;q11) creating a truncated chromosome 22 known as Philadelphia chromosome is present in majority of patients with CML [129]. Depending on the precise location of the fusion, different fusion proteins are encoded. Keeping this in mind and the absence of costimulatory molecules on CML cells, it is improbable that the association of HLA-A*03 is due to its efficiency in presenting fusion proteins and its ability to induce an effective immune response [78]. However, this allele is in with the C282Y mutation of the hemochromatosis gene, a susceptibility marker for CML [78].

In some cases, an optimal immune response is dependent on optimal Ag presentation by both HLA classes and the presence of certain alleles in non-HLA genes. An absence of one of these optimal alleles may result in anergy and immune escape. In some populations, these alleles might be in LD in form of an unknown extended haplotype, while in other populations this haplotype might be absent [57]. One of such associations has been reported between cervical squamous cell carcinoma and multi-locus haplotype of B*4402-Cw*0501-DRB1*0401-DQB1*0301 [57].

Cytokines are a group of soluble regulatory factors by which the immune system controls and modulates different activities of its cells. Each cytokine triggers certain cascade of events in their target cells by binding to their receptors and activating intracellular signal transduction pathway [14, 20]. Cytokine network is responsible for coordination of effector actions of different elements of the immune system, as well as the differentiation and proliferation of different immune cells. In addition, secretion of antibodies and inflammation is tightly regulated by complex interaction between these cytokines [13, 23, 26].

Chronic inflammation, by inducing chronic tissue damage and compensatory cell proliferation, is a considered a major promoter of malignant transformations. As an example, nitric oxide, produced during inflammation, might damage DNA structure in different tumor suppressor genes and oncogenes [130]. Therefore, any dysregulation in cytokine network can result in excessive production of tumor-inducing factors, DNA damage, angiogenesis, and dysplasia and consequent development of various inflammatory diseases including different cancers [26, 131]. Cytokine network is a determinant factor in the development of metastasis and natural history of cancers [26]. In some cancers, malignant cells can manipulate cytokine network in order to escape immunosurveillance or promote their own proliferation [130, 132]. In addition, cytokine network can influence the outcome and toxicity of different immunotherapy methods [13, 20, 133]. Several cancers including hepatocellular carcinoma (HCC), oral squamous cell carcinoma, melanoma, the gastric, pancreatic, and prostate cancer were associated with high levels of certain proinflammatory or antiinflammatory cytokines [26].

Cytokine levels are not the same in all individuals. Interindividual differences in cytokine levels in both baseline and stimulated phases are a result of both genetic and environmental factors [133]. Since there is no an intracellular storage for cytokines, their secretion is dependent on the transcriptional and translational rates of their genes [14, 26]. Not surprisingly, genes responsible for encoding cytokines and their receptors are relatively polymorphic [13, 20, 23]. Several polymorphisms in their gene can affect their expression, structure, and activity [20, 23, 26, 130, 134]. Most of these polymorphisms are in non-coding regions including promoter or intronic sequences and exonic regions are usually highly conserved [13, 14]. So far, numerous genetic association studies have been suggested as associations of these SNPs with various cancers in different populations. However, results of such studies were often inconsistent, and the reported associations varied not only in different populations but also in different cancers and even in their different subtypes [131]. Therefore, a meta-analysis of these studies can show some more conclusive evidence of these associations.

In addition to polymorphisms of cytokine genes, there are other polymorphic elements such as various transcription factors and cytokine-specific receptors which are involved in actions of cytokine network [20, 26]. For instance, polymorphisms in the NF-κB nuclear factor-kappa B gene, one of the most important transcription factors, can result in extensive changes in the cytokine network by altering transcription of TNF-a, IL-1, IL-6, and IL-8 [20]. Although the exact roles of these polymorphisms in tumor immunology are less clear, the relevance of this role is becoming more and more apparent in recent years [20].

IL-1α and IL-1β and their antagonist IL-1Ra are members of this superfamily with pleiotropic effects on inflammation, immunity, and hemopoietic system. High levels of IL-1 are found in tumor sites, however IL-1 family plays an ambivalent role in tumor immunity. IL-1 induces cytokine secretion from T cells to potentiate the differentiation and function of immunosurveillance cells. On the other hand, IL-1 induces the expression of adhesion molecules, matrix metalloproteinases, growth factors, and angiogenic factors and promotes invasiveness and metastasis of malignant cells [135, 136].

Interleukin-4 (IL-4) is a pleiotropic cytokine with major roles in regulation of humoral immunity by its various effects on production of several other cytokines and dedifferentiation of B cells and promoting expression of class II MHC Ags [26, 130]. It also has potent antitumor activity against various tumors by its inhibitory effect on the growth of tumor cells and its growth stimulatory effect on lymphocytes [153, 154].