2.1. Head and Neck Cancer

Head and neck cancers (HNCs) are neoplasms that most commonly originate from squamous cells that make up the epidermis that line the mucosal surfaces inside the head and neck such as the oral cavity, the larynx and pharynx, as well as the nasal and paranasal spaces. Since these tissues are exposed to external factors, the main risk factors have been identified to be tobacco smoke, alcohol, and environmental pollution. In addition, pathogens such as human papillomavirus (HPA) are often causative for this type of cancer as well [21]. The main treatment of HNC remains one that is based on traditional approaches such as surgery, radiation therapy, and chemotherapy, often combined. However, in the last decade, several specific genetic targets have been identified that allow for targeted immunotherapeutic approaches [22]. This search for specific biomarkers has led to the identification of circulating antibodies specifically recognizing the cancer-testis antigen SP17, which was found to be present in 31% of serum samples taken from patients and not present in serum from the healthy control group [23]. The authors compared it to previously reported anti-p53 auto-antibodies, which were found in only 25% of HNC patients [24]. Most importantly, the SP17 antibodies were detected in patients with regionally confined as well as metastasized disease, making it a promising candidate for diagnostic application, especially when combined with additional biomarkers such as the “classical” cytokines and chemokines (IFN-γ, IL-13, MIP-1β, IP-10) [25].

2.2. Central Nervous System Cancers

Central nervous system (CNS) tumors are not easily diagnosed because of the reduced symptomatic manifestations. Moreover, the procedure to obtain a biopsy of suspect masses is a cumbersome, risky, and sometimes impossible procedure. Glioma and meningioma are the most common forms of CNS tumors [26]. O-6-methylguanine-DNA methylase (MGMT) is a biomarker of resistance to chemotherapy in glioma patients. MGMT is involved in DNA repair, and its overexpression and methylation state are routinely investigated in glioma patients to predict drug resistance. A peptide microarray study was conducted to evaluate the production of auto-antibodies against MGMT, and the authors found a good correlation between the presence of the MGMT-2 peptide auto-antibodies and the risk of chemotherapy resistance and disease recurrence [27]. Another study evaluated the proteome signature of meningiomas using human proteome arrays. The study results showed that several proteins overexpressed in grade I and grade II meningiomas are known targets of auto-antibodies, such as IGHG4, CRYM, EFCAB2, STAT6, HDAC7A, and CCNB1 [28]. In a different study, circulating anti-pituitary (APA) and anti-hypothalamus (AHA) antibodies were evaluated in children with CNS tumors. Detectable levels of APA and AHA antibodies were found in patients, but not in healthy controls. In particular, the presence of APA and/or AHA was elevated in patients diagnosed with germinomas, gliomas, and craniopharyngiomas. A similar increase in APA and/or AHA was described in autoimmune pituitary conditions [29].

2.3. Gastrointestinal Tumors

Gastrointestinal cancers are among the most frequent cancers, and they are the third leading cause of cancer-related death worldwide, with an estimation of growing incidence in the next few years. Diffuse gastrointestinal cancers are mainly related to abnormal expression or mutations of E-cadherin, a cell–cell adhesion protein. Several studies have analyzed auto-antibodies’ correlation with gastrointestinal cancer. The most frequently tested and detected are anti-p53 auto-antibodies, but their abundance does not always correlate with gastrointestinal cancer prognosis, stage, or grade, with some exceptions [30]. Adding anti-p53 antibodies to conventional markers significantly improved the overall detection rates of esophageal and colorectal cancers [31]. Stage-related auto-antibody abundance was described for the antigen NY-ESO-1, detectable in significant concentrations in late-stage patients for which the persistence of auto-antibodies is related to poor prognosis. Similarly, anti-AEG-1 auto-antibodies were also mostly found in late-stage patients [32]. Good sensitivity but poor correlation with tumor stage was described for anti-CTAG2, anti-DDX53, anti-MAGEC1, anti-MAGEA3, and GPR78 [32].

2.4. Thyroid Cancer

Differentiated thyroid carcinoma (DTC) is the most frequent form of endocrine tumor with a good survival rate and low recurrence in previously treated patients. DTC treatment often consists of total or partial surgical removal of the thyroid, followed by iodine-ablation ^[33][67]. Cancer clearance and recurrence in DTC is usually monitored by measurement of thyroglobulin (Tg) levels; complete ablation of cancer cells after thyroid removal results in an undetectable level of Tg, and therefore an increase in Tg level indicates only partial cancer removal or cancer recurrence ^[33][34][67,68]. An important aspect in the measurement of Tg levels in DTC patients is the presence of auto-antibodies produced against Tg. Anti-Tg antibodies appear in 10% of normal population and in 15–30% of DTC patients ^[35][69]. Patients with a high level of anti-Tg at diagnosis have a higher probability of disease recurrence. Moreover, anti-Tg levels can be monitored to evaluate disease relapse risk; in particular, stable or increasing concentrations of anti-Tg during DTC follow-up is significantly related with disease persistence and recurrence. Conversely, a decrease in anti-Tg after surgery is a sign of good prognosis ^[34][35][68,69].

2.5. Lung Cancer

The majority of lung cancer (LC) cases are diagnosed at advanced stages, primarily because earlier stages of the disease are either asymptomatic or symptoms may be attributed to other causes such as infection or long-term effects induced by smoking. Auto-antibodies are proving to be a useful tool for early detection. Most studies in this field focus on two classes: either general LC or non-small cell lung cancer (NSCLC).

2.6. Breast Cancer

Since very early lesions of the breast are undetectable by mammography or ultrasound scan, auto-antibodies represent a promising tool to allow for early diagnosis and monitoring of tumor progression from an in situ to an invasive or metastatic phenotype ^[36][94].

HSP60 is one of the breast cancer antigens that stimulate auto-antibody production, and in one study was detected in 31% cases of early stage breast cancer and in 32.6% cases of ductal carcinoma in situ (DCIS) compared to only 4.5% detection in healthy control patients ^[37][95]. In a cohort of women diagnosed with triple-negative breast cancer, a panel of antigens was detected after panning the patient sera with MDA-MB-231 cell line lysate. Out of the detected proteins, the highest score was for PI3K and p53 ^[38][96]. On the basis of similar evidence, researchers proposed an immunodiagnostic model for the prediction of breast cancer versus benign lesions and control ^[39][97]. Another breast cancer auto-antigen is mitochondrial nuclear retrograde regulator 1 (MNRR1), a mitochondrial protein that regulates multiple genes that function in cell migration and cancer metastasis and that is more highly expressed in cell lines derived from aggressive tumors ^[40][98]. Serum p53 auto-antibodies have also been shown to be associated with aggressiveness of breast cancer ^[41][99].

Gu et al. ^[42][100] have shown that anti-cancer antibodies produced by B cells can mediate recruitment of cancer cells derived from the primary breast cancer tumor into draining lymph nodes, thus contributing to the dissemination of the disease. In particular, this pro-metastatic effect was mediated by binding of pathogenic IgGs to glycosylated heat shock protein family A member 4 (HSPA4), a candidate tumor antigen of the HSP70 family. This cell-surface receptor can activate a signaling cascade that is involved in the expression of the CXCR4 ligand stromal-derived factor 1α (SDF1α) in lymph node stromal cells. In patients with breast cancer, high tumor expression of HSPA4 and elevated serum levels of anti-HSPA4 IgG were found to be associated with lymph node metastasis and poor prognosis ^[42][100].

2.7. Adrenocortical Carcinoma

Adrenocortical carcinoma (ACC) is a rare, but aggressive endocrine malignancy, which often has a negative prognosis especially when diagnosed at an advanced stage. Among the determinants of malignant behavior of this tumor, a critical role is played by molecules modulating cell death and resistance to chemotherapeutic agents. One of these molecules is the steroidogenic factor-1 (SF-1) target gene FATE1, encoding for a protein localized at the interface between mitochondria and endoplasmic reticulum, where it regulates Ca²⁺-dependent and mitotane-induced apoptosis in ACC cells by modulating the distance between the two organelles ^[43][101]. FATE1 is expressed at high levels in about 40% of adult ACC and its expression is significantly and inversely correlated with patients’ overall survival. Additionally, FATE1 could be relevant also in other tumors as it has been reported that its silencing increased sensitivity of the NCI-H1155 NSLC cell line to paclitaxel and reduced the viability of a variety of other cancer cell lines. Moreover, circulating antibodies directed against this protein were detected in 3 out of 41 (7.3%) and 4 out of 52 (7.7%) patients with hepatocellular carcinoma in two different studies. Patients with adrenocortical tumors had high tumor FATE1 mRNA expression levels and could mount an immune response against FATE1, as shown by the widespread presence of circulating antibodies directed against this cancer-testis antigen. This is associated with high steroidogenic gene expression, immune cell depletion, and a worse prognosis. High steroid production by FATE1-expressing tumors is likely to create an unfavorable environment for immune cell infiltration and local response against this antigen. On the other hand, FATE1 expression in the most aggressive group of ACC could open new perspectives for immunotherapy using vaccination against this and other cancer neoantigens ^[43][101].

2.8. Ovarian Cancer

The overall five-year survival rate for ovarian cancer is below 30%, as over 70% of patients are diagnosed with stages III or IV disease ^[44][102]. However, subjects diagnosed with localized disease have a survival rate of 75–90%. There is therefore a need to identify biomarkers that have utility for ovarian cancer early detection. At present, cancer antigen 125 (CA125) is the most investigated early detection marker for ovarian cancer and, because the protein detection in circulation has limited sensitivity, tumor-associated auto-antibodies may improve on the performance of CA125 alone. A study by Fortner et al. ^[44][102] demonstrated that, when circulating CA125 levels are examined in the context of anti-CA125 antibodies, their ability for the early detection of ovarian cancer may be improved, especially among women with higher anti-CA125 antibody levels. In these early detected cases, lower circulating CA125 levels were also observed among women with higher anti-CA125 antibody levels, consistent with the hypothesis that higher antibody levels may mask detection of circulating antigen. Circulating immune complexes (CIC) to CA125 have been identified, and potential interference with conventional assays for CA125 has been demonstrated.

2.9. Cervical Cancer

Thus far, auto-antibodies recognizing two different antigens have been identified in cervical cancer—one is the tetra saccharide CA19-9, which strongly increases in patients with more advanced cervical cancer ^[45][113], and one is GAPDH ^[46][114]. Interestingly, in this latter study, Xu and colleagues ^[46][114] found a negative correlation between the level of circulating auto-antibodies and the severity of the cervical lesions. The authors showed that lower anti-GAPDH IgG levels could discriminate between normal cells and more progressive stages of cervical lesions that often lead to cervical cancer (e.g., normal vs. cervical intraepithelial neoplasia (CIN) II and III). This negative correlation of the GAPDH antibody level with different stages of cervical lesion can function as an important biomarker in cervical cancer, especially when combined with additional markers, such as the presence of antibodies to the human papillomavirus (HPV), a pathogen found in nearly all cervical cancer patients ^[47][115].

A combination of ELISAs for anti-CA15-3, anti-CEA, and anti-CA19-9 reliably discriminated CINs from normal cases, and cancer from normal cases, suggesting that this combination assay could be useful for primary screening of cervical cancer ^[45][113].

2.10. Bladder Cancer

While bladder cancer (BC) is often not perceived as such, it is ranked sixth in absolute incidence rate for men worldwide, while only seventeenth for women ^[48][116]. The USA and Europe see the highest incidence rates often related to exposure risk factors such as cigarette smoking, causative of up to 50% of all BC diagnosed as well as some environmental chemical carcinogens ^[49][117]. Most BC patients are initially diagnosed with a better-treatable form defined as non-muscle-invasive disease (NMIBC). Several cancer progression subclassifications (using three stages and two grades) have been introduced to define the chance of recurrence, which, in more progressive cases, is very high (50–70%) and will eventually lead to a muscle-invasive disease (MIBC) variant that is associated with a very poor prognosis ^[49][117]. For this reason, adequate biomarkers are also needed in bladder cancer in order to identify the aggressive high-grade MIBC. Minami and colleagues ^[50][118] found that the expression levels of an important serine/threonine-specific protein phosphatase, PPP1CA, involved in signal transduction, apoptosis, protein synthesis, and cell-cycle regulation, was associated with poorer prognosis, and the authors subsequently investigated IgG serum levels raised against this protein. Serum levels of anti-PPP1CA IgG were found to be higher in BC patients than in healthy individuals, with a specificity of 64.2% and a sensitivity of 65.7%, and were associated with muscular invasion, a higher tumor grade, and poorer prognosis, making it a potential valuable diagnostic and prognostic marker. Besides a role as diagnostic biomarker, understanding the role PP1Ca plays specifically in cancer progression could lead to new therapeutic interventions for this type of cancer ^[51][119].

2.11. Prostate Cancer

In a study by Tan et al. ^[52][120], the proteins SPARC and Fetuin-A were selected for analysis of auto-antibodies as they were shown to be highly expressed at late stages of prostate cancer. Sera from 117 Caucasian American (CA) and 111 African American (AA) prostate cancer patients with Gleason grades 6–10, and healthy controls (CA, n = 52; AA, n = 45) were analyzed in addition to sera from a biopsy cohort (n = 99). The specificity of auto-antibodies against the respective target proteins was confirmed by immunoblot analysis. Both SPARC and Fetuin-A antibodies were detected in the sera, with significantly lower levels in both CA and AA prostate cancer patients compared to healthy controls. The range of auto-antibodies reactivity to SPARC and Fetuin-A was similar in both CA and AA prostate cancer patients, indicating a similar behavior also across ethnic groups ^[52][120].

2.12. Testicular Seminoma

Testicular seminoma accounts for 40% of testicular cancers and originates from the germinative components of testicular epithelium. Testicular seminoma is highly treatable and has a good prognosis. Diagnosis is mainly made by ultrasound investigation. Often, the onset of a testicular disease is accompanied by paraneoplastic encephalitis and, indeed, almost 20% of patients with paraneoplastic limbic encephalitis have concomitant testicular seminoma ^[53][121]. Several studies have shown the presence of serum auto-antibodies in patients with paraneoplastic encephalitis, of which, presently, six are recognized as indicative of an ongoing paraneoplastic encephalitis ^[54][122]. In particular, high levels of anti-Ta antibodies in men are associated with testicular seminoma diagnosis. Anti-Ta antibodies react with the paraneoplastic protein PNMA2 (former antibody name: anti-Ma2) ^[55][123]. In many cases, anti-Ta antibodies can be detected several months before diagnosis ^[56][124]. Mandel-Brem ^[57][125] and, more recently, Maudes et al. ^[58][126] have described the association of detection of Kelch-like protein 11 auto-antibodies to testicular seminoma in patients with paraneoplastic encephalitis. The presence of elevated concentrations of serum anti-Kelch-like protein 11 auto-antibodies, alone or in combination with anti-Ma auto-antibodies, was associated with a poor response to treatment of testicular seminoma ^[58][126].

2.13. Lymphomas

Non-Hodgkin’s lymphoma (NHL) is the most common form of lymphoma, a cancer that affects the lymphatic system. A 2012 study investigating the role of the immune system in the progression of NHL found significant levels of circulating auto-antibodies in 150 patients, 50% of which had large B cell lymphoma. The patients were either newly diagnosed, received treatment (i.e., chemotherapy), or were disease-free during follow up. In total, 84% of the patients had one or more auto-antibodies. The antigens identified were Jo-1 and Jo-3, single strand DNA (ssDNA), perinuclear anti-neutrophil cytoplasmic (p-ANCA), antinuclear (ANA), and rheumatoid factor (RF). Patients with newly diagnosed NHL had significantly higher levels of anti Scl-70, anti Jo-1, and RF compared to other patients, indicating the auto-antibodies could not only be used for diagnosis, but could potentially also help with staging ^[59][127]. Interestingly, the detected levels of anti-double strand DNA (dsDNA) and anti-ssDNA were relatively low compared to another study by Swissa et al., who showed high levels in 23.8% NHL patients as well as anti-RNP and anti-SM antibodies that were not detected in most of the control group ^[60][128]. An extensive investigation into auto-antibodies and recognized neoantigens in human mantle cell lymphomas (MCLs), a rare type of non-Hodgkin lymphoma, was undertaken by Khodadoust et al. ^[61][129] using a very thorough genomic and proteomic approach combining MHC isolation, peptide identification, and exome sequencing ^[61][129]. However, since no control group of healthy patients was included, no information could be given on diagnostic sensitivity or specificity using these auto-antibody levels. Interestingly, all identified neoantigens were peptides exclusively derived from the lymphoma immunoglobulin heavy or light chain variable regions and, surprisingly, no neoantigens were identified from non-immunoglobulin somatically mutated genes. The study showed that many of the genes presented by MHC were shared between the 17 patients used for the study. However, the peptides were generally unique to each patient except for those with similar MHC-I and/or MHC-II alleles. The authors concluded that, although they could not completely exclude a certain experimental bias towards more abundant epitopes that might have not allowed identification of very rare epitopes, the possibility to identify lymphoma-specific CD4 neoantigens and their use to select and expand endogenous T cells could present an effective lymphoma immunotherapy.

2.14. Melanoma

Different from lymphomas and many other cancers, melanomas are often characterized by a very high mutational load, which increases the likelihood of this type of tumor generating neoantigens. In melanomas, the prognostic value of the mutational load has been investigated and was found to be associated with the clinical benefits of immunotherapy such as the adoptive T cell therapy (ACT), which has shown remarkable results in clinical trials in certain patient groups while failing in others ^[62][130]. A comprehensive genomic analysis of melanoma tumor samples by Lauss et al. ^[63][131] showed a strong positive correlation between a high mutational load and improved clinical outcome following ACT. This also aligns well with the observation that tumors with a high mutational load respond better to treatment with immune checkpoint inhibitors in melanoma and lung cancer ^[64][65][132,133], and is indicative of an increased production of neoantigens. From these and other studies it has also become clear that the presentation of tumor antigens is dependent on the MHC class I antigen processing pathway and subsequent recognition by CD8⁺ T cells. Loss of MHC I antigen presentation was found in many advanced-stage melanomas while high-level expression of MHC class I antigen processing machinery (APM) genes was associated with clinical benefits to immune therapies. When combined, the studies on predictive outcome from ACT seems to indicate that the neoantigen load, predicted by the measured mutational load, can function as an independent predictor for survival after ACT. Two more important melanoma characteristics that were found to be strongly associated with improved outcome from immune therapy were a low-proliferative status and the upregulation of genes responsible for antigen presentation by the cells ^[64][65][132,133].

Due to the high mutational load and consequential formation of neoantigens, circulating auto-antibodies could function as important biomarkers for early detection, incredibly crucial for treatment of this disease. Using protein microarrays, melanoma patient sera was compared with healthy control sera. Antibodies in the patient sera (124 patients) recognized a total of 748 antigens, of which 139 stood out, giving a mean specificity of 97% ^[66][134]. Of the 139 antigens, 20 were not reactive with healthy control serum. Interestingly, most of the seroreactive proteins were found to be intracellular (101/139), many of which were nuclear (88/139). Many of the top 139 identified reactive antigens belong to common cancer pathways involved in apoptosis, cell cycle, p53 signaling, MAPK signaling, and immune response. The authors concluded by stating that the detection of combinations of auto-antibodies gave superior sensitivity and specificity, and for their cohort, a set of 10 auto-antibodies gave 79% sensitivity and 84% specificity, much higher than can be accomplished with existing individual biomarkers. A further combination of 10 autoantibody biomarkers (ZBTB7B, PRKCH, p53TP53, PCTK1, PQBP1, UBE2V1, IRF4, MAPK8_tv2, MSN, and TPM1) displayed a sensitivity of 79% and specificity of 84% for primary melanoma detection ^[66][134].

With high-throughput screening technologies, it is easy to envision that screening for larger sets of auto-antibodies will further increase the accuracy of diagnostic testing.

2.15. Angiosarcoma

Angiosarcoma is a rare soft tissue sarcoma, accounting for less than 2% of total soft tissue sarcomas. Non-specific symptomatology and manifestation, together with the high aggressiveness and propensity of spreading of this type of cancer, make angiosarcomas challenging both for diagnosis and for treatment ^[67][135]. Angiosarcoma is often positive for over-expression of VEGF and its receptor (VEGFR), both of which can be the possible target for therapy based on anti-angiogenic agents ^[67][135]. Diagnosis of the symptoms and location of angiosarcoma is quite challenging, but recently an association between angiosarcomas and anti-p53 serum auto-antibody levels has been proved statistically relevant and can be exploited both for early diagnosis and for disease staging ^[68][136].

2.16. Antibody-Mediated Paraneoplastic Syndromes

There are several paraneoplastic syndromes associated to auto-antibodies. Many of them are due to onco-neural auto-antibodies ^[69][137], whose molecular and pathogenic relationship with the original cancer is often difficult to identify. Anti-Ri antibodies ^[70][138], breast cancer anti-phospholipids antibodies ^[71][72][139,140], cerebellar degeneration ^[73][74][141,142], glomerular diseases ^[75][143], and lung cancer ^[76][144] are all conditions which still require a proper pathogenic explanation.

To summarize, Table 1 presents the antigens mentioned in the above paragraphs in alphabetical order and Table 2 lists the main auto-antibody combinations that seem promising in early cancer detection or in the prediction of cancer prognosis. Where available, sensitivity and specificity were also included in dedicated columns.