More or less

Abstract

Current laboratory investigation of patients with suspected immunodeficiency or immune-mediated disease includes extensive analysis of lymphocyte subsets, T-cell receptor use, antibody profiles, cytokine profiles and genetic polymorphisms of relevant genes, all in all: big data. Clinical immunology clearly has entered the omics era, generating more and more data.

Keywords:

• antibody response
• bioinformatics
• clinical immunology
• data mining
• phenotyping
• vaccination

Cosmas and Damian, the first transplantation surgeons in recorded history, did not perform any diagnostic testing (such as HLA typing, immune status, infectious status) before they grafted a leg of an Ethiopian servant on his Venetian master [1]. Nevertheless, the transplantation was successful and Cosmas and Damian became saints.

Seventeen centuries later, clinical immunological laboratory diagnostics were in place when Dr. Ogden Bruton described the first patient with agammaglobulinemia in 1952 [2]. The single laboratory test he performed was serum electrophoresis in agar. The strongly reduced intensity of the proteins migrating in the gamma region of the electropherogram was sufficient to make the diagnosis. During the next 60 years, the rapid increase in the knowledgebase of the immune system and the accompanying developments in technology have revolutionized the immunological laboratory.

Before the discovery of the monoclonal antibody technology by Kohler and Milstein [3], human T lymphocytes were characterized and enumerated based on their ability to bind sheep red blood cells (rosette forming cells). B lymphocytes were detected based on their binding of FITC-labeled (polyclonal goat) anti-human immunoglobulin antibodies. T and B lymphocytes were counted manually by (immunofluorescence) microscopy. The remainder of the lymphocytes, which were neither T- nor B-cells, were termed 0-cells, so peripheral blood lymphocytes could be divided into three subsets. Starting with monoclonal antibodies against CD3, CD4 and CD8 antigens, T lymphocytes could be subdivided into helper/inducer cells (CD4⁺) and cytotoxic/suppressor cells (CD8⁺). The almost parallel development of flow cytometry as a high-throughput analysis technique allowed for implementation of lymphocyte subset analysis in immune status protocols. Currently, with over 350 reagents against cytoplasmic and cell surface proteins available and 6–8 color flow cytometry, up to 95 and even more lymphocyte subsets are reported by clinical immunological laboratories [4].

These, and other big data generated by immunological laboratories on a daily basis, do not automatically make it easier to provide advice on the immune status of a patient with an increased frequency of infections, or a patient with hypersensitive reactions to compounds in the environment, or a patient with joint complaints. The need for a bioinformatics approach, which makes sense of the data, is obvious. The case of specific polysaccharide antibody deficiency (SPAD) will be given as an example. SPAD is defined as the specific inability to produce antibodies to polysaccharide antigens, while the antibody response to protein antigens is normal [5]. In order to confirm the diagnosis in a patient with recurrent respiratory tract infections with polysaccharide-encapsulated bacteria (such as Streptococcus pneumoniae), patients are vaccinated with a polysaccharide vaccine containing the capsular polysaccharides of 23 most prevalent serotypes and the subsequent antibody response is measured. According to the definition of SPAD, the diagnosis can be made when the patient has an insufficient response to more than 70% of the tested serotypes [6,7]. In the ELISA-era, antibody concentrations were measured by solid-phase immunoassay, in which the antigen is bound to the polystyrene surface of a microtiter plate. Although the technique is very robust and reliable, the downside is that it allows to measure only one serotype at a time. Measuring more serotypes means running more ELISA plates, which requires more blood and more technician time. For that reason, most laboratories restricted their analysis to four to six serotypes at best. The introduction of multiplex immunoassays (Luminex) has changed this type of diagnostics. Nowadays, only a single sample is needed to measure antibody levels against 25 different pneumococcal serotypes [8]. A listing of those 25 antibody titers cannot constitute a laboratory report, because it would create more questions than answers. There is a clear need to develop algorithms that can incorporate the individual antibody titers into a qualitative estimate of responsiveness.

In fields such as oncology, the systems approach has been taken in gene expression profiling and tissue microarrays. A variety of data mining techniques, including principal component analysis, clustering, dendograms and so on is used to optimize diagnostics, prognosis and personalized treatment. More and more data are now also generated in the immunology laboratory for establishing the composition and function of the immune system [9–12]. In order to be useful for diagnosis and treatment of individual patients, today’s challenge is to interpret those data and condense it into a useful advice for physician and patients.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.