Current applications of antibody microarrays

Signalling pathway related

Changes in signalling pathways is a hallmark of many disease states, including cancer [23, 24], diabetes [25, 26] and neurodegenerative disorders [27]. Building antibody arrays using capture reagents targeting proteins in the signalling cascade of interest has enabled researchers to investigate changes in protein profiles and modifications in signalling pathways in normal biological processes and disease states. Researchers have applied this concept to a range of biological sample materials and preparations, including cell lysates [28], tissue extracts [29], and plasma [30].

Calbindin-D28 k (CB), an important calcium-binding protein that acts as a calcium buffer and is found to be expressed at lower levels in the brains of mice and humans with Alzheimer’s disease but it is not known if these changes contribute to AD-related dysfunction (AD). Kook et al. generated a CB-deficient Alzheimer’s transgenic mouse (CBKOTg) to investigate in CB’s contribution to signaling pathways in AD [23]. Utilizing antibody microarrays to examine mouse brain tissue, they identified significant alterations in cell death pathways, synaptic transmission and MAPK signalling pathways upon CB deletion. These findings were verified by immunohistochemistry which showed increased apoptotic markers and increased neuronal death, demonstrating that antibody microarrays can provide novel information, enhancing our understanding of the role CB and its implication in AD pathophysiology.

Antibody microarrays have been used to uncover the physiological role in highly conserved proteins, including elongation factor 4. Gao et al. dissected the mechanism of quality-control factor mitochondrial Elongation factor 4 (mtEF4) in translation using a phospho-explorer antibody microarray [31]. By comparing cell lysates between mtEF4 knock-out and WT mice testis tissues, they found that the fold enrichment of mTOR was the highest among different signalling pathways. The authors showed that with mtEF4 deletion, the major feedback signal from the somatic cytoplasm is mTOR upregulation. This was accompanied by an increased cytoplasmic translation, thus indicating that mTOR plays a crucial role as a downstream effector compensating for mitochondrial translation deficiency [32]. Furthermore, the study linked previously undescribed cross-talk between mtEF4-dependent quality control in the mitochondria and the mTOR pathway in the cytoplasm.

Drug mechanism

With the help of antibody microarray assays, researchers can investigate drug mechanisms in a systematic and efficient manner. Among many possible applications, tumor-induced angiogenesis has been studied as it plays a pivotal role in cancer progression [33]. P11, a novel peptide ligand containing a PDZ-binding motif (Ser-Asp-Val) with high affinity to integrin α_vβ_3, was identified from a hexapeptide library (PS-SPCL) using an antibody microarray [34]. The pharmacological mechanism of P11, was elucidated using a specifically designed pharmacoproteomic microarray approach containing 48 cancer related antibodies [35]. The study revealed that P11 inhibits bFGF-induced human umbilical vein endothelial cell proliferation via mitogen-activated protein kinase and extracellular-signal regulated kinase inhibition. In addition, the microarray revealed P11 caused the upregulation of apoptotic marker p53, resulted in apoptosis induction via activation of the caspases system, which indicates P11 may play a key role in preventing tumor progression. The link between P11 and p53 presents clinical and basic cancer researchers will help to elucidate further clues about the role of P11 and its inhibitory target.

Lovastatin, a natural product derived from Aspergillus terreus or Monascus ruber, has been widely used as a cholesterol-lowing drug in the clinic. Yang et al. have used an antibody microarray containing 656 antibodies with a focus on different functional cell pathways, which suggested that lovastatin also has anti-cancer properties, through poorly defined mechanisms. There is a strong need to find effective cancer treatments and to understand their mechanisms of action. Using a microarray, breast cancer cells were studied under hypoxic conditions to investigate the molecular mechanisms through Lovastin exherts its effect. They showed 17 up-regulated proteins and 20 down-regulated proteins for lovastatin treated breast cancer cells, compared to the control, and the results were subsequently validated by real-time PCR [36]. The protein signature included proteins involved in apoptosis, cell proliferation and tumour metastasis, linking modulation of these pathways to the pharmacological action of lovastin.

The type 2 diabetes drug metformin, has also proposed to have anti-cancer properties. Lee et al. investigated the potency of novel metformin derivatives using an antibody microarray built to study cell cycle-related molecules. The authors showed that metformin-butyrate (MFB) had better anti-tumor efficacies than metformin-HCl, through greater neoplastic activity and better efficiency at impairing cell cycle progression during S and G2/M phases in addition to appearing to have preferential cytotoxic effects on breast cancer stem cell populations [37]. These findings highlight how antibody arrays can provide proteomic evidence supporting development of drugs with better efficacy.

For basic biology research, antibody microarrays can serve in detecting signalling proteins and proteins associated to a particular phenotype. Arrays with parallel protein detection create an opportunity to investigate multiple signalling pathways in a single experiment, providing insights into novel mechanisms, such as the study of disease progression, drug interaction, and response to infections.

Autoimmune diseases

The diagnosis of systemic lupus erythematosus (SLE) is challenging due to its heterogeneous clinical presentation and the lack of robust biomarkers to distinguish it from active and inactive disease as well as from other autoimmune diseases. Lin et al. [38] performed a DotScan™ antibody microarray screening of peripheral blood mononuclear cells from 60 SLE patients of varying disease activity, 25 rheumatoid arthritis patients, 28 other autoimmune disease samples, and 24 healthy controls. The antibody microarray profiles could distinguish active SLE patients from healthy controls. Using a leukocyte capture array improved the discriminative ability of conventional SLE diagnostics, by verifying serum anti-dsDNA, complements C3 and C4, the microarrays increased the capability of discrimination of semi-active and active SLE, information which will assist better disease management.

In another study, Carlsson et al. [39] constructed 135 human recombinant single-chain fragment variables (scFv) targeting immune proteins to build an in-house antibody microarray. In this study, they examined patients with systemic sclerosis (SSc), SLE and 15 healthy volunteers. SSc is an autoimmune disorder affecting connecting tissue, which can be challenging to differentially diagnose from SLE. The array identified 40 differentially expressed proteins creating a candidate proteome signature to delineate SLE and its severity from SSc. This protein signature was a better disease classification than single or even combinations of conventional clinical parameters, including, ANA, anti-DNA, SLEDAI-2 k, C1q, C3, C4 and CRP, illustrating the potential use for antibody microarrays to create new disease-signatures which will add clinical value to disease management.

Infectious diseases

Infections activate the immune system differently, and these mechanisms can be investigated using antibody arrays on a proteome-wide level. Helicobacter pylori is a pathogen that colonizes roughly one half of the world’s population which causes chronic gastritis. Sukri et al. [40] employed the DotScan™ antibody microarray to determine the tolerance of immune system toward tumor cells in gastric cancer using 144 CD antibodies to profile the distribution of CD markers between Helicobacter pylori infected and un-infected gastric adenocarcinoma cells. Interestingly, they found gastric adenocarcinoma cell line AGS infected by cagA + H. pylori showed increased CD27 expression, which is essential for maintenance of T cell population, and increased CD markers were also detected in H. pylori-infected gastric cancer patients. This study suggests not only the tolerance of the immune system toward gastric cancer, but also the immune response variations exploited by different H. pylori strains.

Furthermore, antibody arrays can be useful in tracking a continuing change in the physiological environment, including disease status and response to therapies and interventions. After liver transplantation, 94% of patients were found with hepatitis C virus (HCV) induced histological damage, with some patients suffering severe disease recurrence [41]. There is an urgent need for predictive biomarkers to identify recurrent disease severity. Using a CD antibody microarray, peripheral blood from patients defined as pre-transplant, early, mid, and late-transplant groups were investigated to predict the severity of HCV recurrence after transplantation [42]. Serial blood samples taken from patients before and after liver transplant enabled tracking of the disease milieu, with its own internal controls. Five CD markers (CD27, CD182, CD260, CD41, and CD34) were significantly increased in severe recurrence compared to mild recurrence. This result shows that antibody arrays may assist in assessing recurrent HCV disease severity after liver transplantation.

Ellmark et al. [43] devised a single framework recombinant antibody (SinFabs) microarray containing 127 different antibodies against immune-regulatory antigens selected from the n-CoDeR library on gastric adenocarcinoma. They elucidated that distinct tumor- as well as infection-associated protein expression signatures, such as IL-9, IL-11, and MCP-4, could be identified from the plasma proteomes and serve as potential biomarkers. These findings may aid to improve the understanding of H-pylori induced cancer and may pave the way for refined diagnostics in the future.

Bead based antibody arrays have been performed to investigate differential protein profiles in plasma in children suffering from malaria and malaria-related complications [44]. From 1000 proteins screened, 41 proteins had differential expression between malaria infected children and community controls. Thirteen further proteins were linked to malaria-disease severity. These findings indicate the involvement of inflammation as well as unbalanced metabolism of glucose in severe forms of the disease and the biggest changes were seen in two of muscle proteins (carbonic anhydrase 3 and creatine kinase) pointed at muscle damage and lesions for children with cerebral malaria. These findings may aid the development of simple tests for classifying malaria-infected children into group showing a higher risk for the severe form of the disease.

Cancer

Antibody microarrays have been applied for cancer research mainly to study cancer progression and candidate proteins that may serve as diagnostic biomarkers [45, 46]. Pancreatic cancer is an aggressive disease with poor prognosis, and disease-specific biomarkers that offer early and accurate diagnosis are in urgent need. Using an in-house developed recombinant antibody microarray platform, Wingren et al. [47] screened sera from 148 patients with pancreatic cancer, chronic pancreatitis, autoimmune pancreatitis (AIP), and healthy controls. They identified a panel of 25 protein targets, including IL-2, IL-11, IL-12, TNF-, which contributed to distinguishing pancreatic cancer from healthy controls. This 25 protein signature exhibits a high diagnostic potential (AUC of 0.88). In addition, the group has recently extended their analysis of their panel to additional sample sets, including lymphoma [48], prostate cancer [49], and breast cancer in cell lysates [50] and plasma [51]. These findings are useful to aid the development of novel and multivariate diagnostics methods.

One of the most important questions in cancer biology, is determining the invasiveness of a tumor. One study developed a customized an antibody microarray platform containing 4096 features to interrogate plasma samples spanning pre-invasive and invasive diseases from a mouse model of pancreatic ductal adenocarcinoma (PDA) [18]. They found a protein signature, comprising of the differential expression of three proteins, ERBB2, TNC and ESR1, which could be used to improve the AUC from 0.86 (95% confidence interval [CI] 0.76–0.96) to 0.97 (95% CI 0.92–1.0) when the PDA marker CA19-9 included.

The majority of newly diagnosed cases of bladder cancer are low-stage, low-grade, non-muscle-invasive [52]. After standard transurethral resection, 50–70% of tumors recur, however 10–30% of the tumors will progress to muscle-invasive disease [53, 54]. Understanding the mechanism of tumor progression could provide more useful information to clinical scopes. Srinivasan et al. studied cell lysates with an antibody microarray containing 810 cancer related antibodies constructed by the Hoheisel lab [55]. The authors built a multivariate classifier containing 20 proteins, which facilitatated the prediction of recurrence with a sensitivity of 80% and a specificity of 100%. Interestingly, they found repression of the TGF-β signaling pathway in recurrent cancer. The signaling factors IFNG, TNF-α, and THBS1 were expressed less and the abundance of the inhibitor MAPK3 (also known as ERK1) was higher, while SMAD2, SMAD3, and SMAD4 were again significantly underrepresented [56]. The data indicated that TGF-β signaling pathway inhibitors may reduce bladder cancer recurrence [57].

Puig-Costa et al. [58] employed antibody microarrays to discover biomarkers for gastric cancer (GC). Notably, the antibody microarray contained antibodies targeting different well-known functional proteins that play a crucial role in cancer progression, including 120 cytokines, 43 angiogenic factors, 41 growth factors, 40 inflammatory factors and 10 metalloproteinases. Ingenuity pathway analysis confirmed that some biomarkers, such as ICAM-1 and angiogenin, indicated high inflammatory response in GC patients. Cellular movement and immune cell trafficking targets such as monocyte chemoattractant protein (MCP)-1 was found to be overrepresented within GC patients. The positive predictive and negative predictive values in this validation cohort were 75% (95% CI 53–90) and 80% (95% CI 56–94), respectively. Finally, antibody microarray analyses of the GC-associated inflammatory proteome identified a 21-protein inflammatory protein-driven gastric cancer signature (INPROGAS) that accurately discriminated GC from noncancerous gastric mucosa, and may provide new leads for the analysis of cancer progression.

In prostate cancer, Schwenk et al. [59] used antibody arrays on suspension bead arrays to compare plasma levels of proteins between different groups in order to find additional biomarkers alongside prostate-specific antigen (PSA). Besides classifying the patients based on PSA, they identified decreased plasma CNDP1 levels and in a subsequent larger studies this was shown to association in more aggressive forms of the disease. A sandwich immunoassay was developed to validate these findings in more than 1200 patients, and the association of the decreased CNDP1 to lymph node metastasis was further elucidated [60]. This demonstrates one of the very few studies that have successfully validated an initial indication through the antibody microarray pipeline and the development of a targeted assay. In a subsequent investigation of cancer cachexia, plasma levels of CNDP1 was found to be decreased in cachexic patients, hence pointing at a metabolic role of changes in CNDP1 levels [61].

In a study of small intestine neuroendocrine tumours, Darmanis et al. [62] investigated the plasma levels of two independent study set (77 and 132 samples) and a targeted bead array for 124 unique proteins. They could achieve classification accuracy of up to 85% using a panel of proteins and concluded with proposing novel candidates for classifying the tumors. Of the shortlisted candidates, such as for IGFBP2 and IGF1, the group performed ELISA assays to confirm the indications. These findings indicate a metabolic association of plasma proteins with cancer and support to consider these targets in future cancer studies.

Neurodegenerative diseases

Analysis of protein expression provides a possibility to extend the current knowledge on neurodegenerative pathophysiology [63]. Recent advances within the field of proteomics have offered the potential to search for novel biomarkers using antibody microarrays [64]. Advancement in protocol development have allowed researchers to study not only blood, but also cerebrospinal fluid (CSF) [69].

Within neuroproteomics, the Nilsson lab has conducted several studies by using antibodies form the Human Protein Atlas [65] together with the suspension bead array assays. Profiling CSF from patients with Multiple Sclerosis (MS), they found GAP43, a cytoplasmic protein involved in the formation, and regeneration of neurons, hence a promising biomarker diseases of the brain [66]. More recently, Remnestål et al. [67] compared protein levels in 441 CSF samples collected from different neurodegenerative disease sample sets as well as CSF collected post-mortem. Among 376 antibodies, the synaptic proteins GAP43 and NRGN were found to be associated to AD patients compared with controls.

Using plasma samples, CSF and brain tissue from patients suffering from MS, a large-scale screening was conducted that started from utilizing 4500 antibodies on bead arrays [68]. A set of proteins was found to be associated with MS subtypes in CSF and plasma. Utilizing some of the candidate antibodies raised against IRF8, IL7, and METTL14 also for immuno-fluorescence analysis of brain tissue showed staining of neurons in proximity of MS lesions. This indicates that antibodies selected from array-based assays for analysis of body fluids can also provide further evidence at the affected tissue.

Lastly, the bead arrays were used to profile plasma from patients suffering from amyotrophic lateral sclerosis (ALS) [69], where 367 ALS patients and 101 controls were analyzed for 278 proteins. The study concludes with proposing neurofilament medium polypeptide (NEFM), solute carrier family 25 (SLC25A20), and regulator of G-protein signalling 18 (RGS18) as valuable proteins because they are involved in processes related to disease pathophysiology, which warrant further validation in independent sample sets.

Flow cytometry

There are further developments and novel applications of antibody microarrays. Sharivkin et al. [103] showed a novel method of combining antibody microarrays with flow cytometry to isolate specific cell types from differentiating stem cell populations. An array of antibodies against cell-surface antigens was printed on a hydrogel coated glass slide. Live cells isolated by different markers using flow cytometry. And then captured on specific antibody spots by interaction between their surface antigens and the printed antibodies. This allowed identifying additional markers that further refined different subpopulation from flow sorting [104].

CRISPR/Cas9

The bacterial clustered regularly interspaced short palindromic repeats (CRISPR)-Cas9 system for genome editing has greatly expanded the toolbox for mammalian genetics, enabling the rapid generation of isogenic cell lines and mice with modified alleles. Wang et al. [105] generated a knockout pool for two human cell lines, and this loss-of-function genetic screening approach was suitable for both positive and negative selection. With 1000’s of genes to manipulate in one experiment, the antibody microarray has a great opportunity to be combined with this pool and the CRISPR/Cas9 technology holds promise to create a whole human proteome knock out library, which had the potential to revolutionise how we study proteins and their interactions in health and disease.

Advance materials

Chinnasamy et al. presented a novel lateral flow microarray-based device using a novel dually labelled gold nanoparticle-strategy for rapid and sensitive detection of clinical serum samples. Each gold nanoparticle was conjugated to an optimized ratio of HRP and anti-IgE, allowing a significant improvement of assay sensitivity as compared to commercially available detection reagents. Also, due to the rapid and simple procedure, inexpensive materials and read-out by means of a consumer flatbed scanner, the presented assay may provide a sensitive and low-cost platform for multiple fast testing which could facilitate antibody microarray in translation medicine area [106].

Antibody microarrays have been established alongside the advancement of materials. Hu et al. [107] developed a hierarchically nanostructured organic–inorganic hybrid substrate, comprising of randomly oriented ZnO nanorods on glass slides with coaxially tethered dense polymer brush, which highly improved the limit of detection (LOD) to 100 fg/mL. Huang et al. [108] have invented a glycol-gold nanoparticle-based antibody microarray, with potential to amplify a signal that could be detected by the naked eye.