In P/P patients, antibodies generated by B cells of the immune system block the function of desmoglein proteins Dsg1 and Dsg3 known to be important in binding together keratinocytes of the skin and mucous membranes, but it is not known how the rogue antibodies are generated by the immune system, how they escape the quality control mechanisms in place that only allow B cells with non-“self” specificities to survive, and why P/P patients are so rare.
New research led by Dr. Aimee Payne in the Department of Dermatology at the University of Pennsylvania (Nature Communications, http://www.nature.com/ncomms/2014/140619/ncomms5167/abs/ncomms5167.html) helps us begin to understand why.
In previous work, Dr. Payne and colleagues have identified antibodies that recognize Dsg1 and Dsg3 (so-called anti-Dsg1 and anti-Dsg3 antibodies) and have also identified regions of those antibodies that are important for the ability of those antibodies to be pathogenic — that is, to recognize their Dsg targets in pemphigus vulgaris (PV) and pemphigus foliaceus (PF) and to disrupt their function. To extend this work and to better understand how PV autoantibodies arise, Dr. Payne and colleagues have performed a similar analysis of PV patients.
PV patients can present as either mucosal-dominant, where only the mucous membranes are affected or as mucocutaneous, affecting both the mucous membranes and the skin. Almost all mucosal-dominant PV patients have anti-Dsg3 autoantibodies, while the mucocutaneous patients have anti-Dsg3 autoantibodies as well as anti-Dsg1 autoantibodies. Since it is thought that Dsg1 and Dsg3 can compensate for each other’s function, the presence of functional Dsg1 in the skin in the presence of anti-Dsg3 autoantibodies can explain why mucosal-dominant patients do not have skin lesions.
The authors first isolated the full antibody repertoire from four different untreated PV patients, all with mucocutaneous disease. They isolated and characterized these in a multistep process that ultimately allowed them to determine the amino acid compositions (by cloning and DNA sequencing methods) of the patient’s PV antibodies. This led to the assignment of six unique antibodies from Patient 1 and five additional unique antibodies from the three other PV patients.
In total, the sequencing efforts identified 21 unique heavy chains among the four patients.
All 11 antibodies could bind to Dsg3 and this was mediated via a domain (called EC1) in Dsg3 that is known to be important for its adhesive interactions, suggesting that anti-Dsg3 autoantibody binding to Dsg3 leads to a direct block in Dsg3 function in keratinocytes (and subsequent skin blistering).
Curiously, not all of the antibodies that the authors identified that bound Dsg3 could cause blistering when added to human skin tissue samples; the VH1-46-containing antibodies did. They determined that these differences in functional effects were due to the inability of the nonpathogenic antibodies to bind to the functional domains of Dsg3.
Even more curiously, the authors found that all four patients had at least one PV antibody that consisted of an identical variable region termed VH1-46. They also found very little change in the VH1-46 amino acid sequence in the patient antibodies compared to the known sequence of VH1-46 that also exists in unaffected patients (considered the “wild-type” or germline sequence).
As noted by the authors, this is a pattern typical of a somatically mutated antibody sequence, meaning that very few changes were generated during the development of the B cells (each with its own single antibody that it makes, see Figure).
They did some additional experiments to define the ability of those amino acid changes to affect the binding to Dsg3. They conclude that VH1-46 autoantibodies in PV are generated during B cell development and require very little mutation to become pathogenic. This suggests that they appear early during the development of the disease and explains their prevalence in all of the patients tested here.
These autoantibodies may not be the most common later on (during full-blown disease), but they may provide a clue to why and how pemphigus arises. The ability of these autoantibodies to escape the quality control machinery at play during B cell development is likely due to the low levels of Dsg3 antigen available that would distinguish these antibodies as “self” antibodies and therefore the ability of the machinery to mark the cells and their rogue autoantibodies for destruction.
These data led the authors to speculate that the five pathogenic (disease-causing) VH1-46 anti-Dsg3 mAbs that they’ve identified in this study are among the earliest autoantibodies formed in PV patients, caused only by how simple they are to generate from germline sequences. They also define a mechanism for how these autoantibodies are made and most importantly, how they are missed by the quality control machinery – all low probability events that likely account for the rarity of PV.