Monitoring B-cell Depletion Therapy
Baffled about B-cell recovery?
In the previous Newsletter, a relatively well-recognized phenomenon of B-cell recovery following rituximab was shown and discussed: widely variable kinetics of B-cell reconstitution, with B-cell counts reaching or exceeding pre-rituximab levels after only a few months in some individuals, while showing no substantial recovery in others. It was suggested that this might be related to the BAFF system.
The BAFF system (see sidebar for an acronym legend) is a critical component in B-cell homeostasis. The seemingly unremitting publication of review articles, editorials and commentaries is indicative of the importance of this system, but also of the steep angle of the BAFF-learning curve. Before going into more detail, it is important to remember that many attributes and characteristics of the BAFF system are based on studies in mice, including mice with genetically altered BAFF components. The BAFF system belongs to the tumor necrosis factor (TNF) superfamily, and can be roughly divided into secreted factors/ligands and receptors. On the ligand side, the main players are: homotrimeric or polymeric BAFF (proteolytically cleaved from the cell surface membrane), APRIL (only secreted form), and BAFF-APRIL (in heterotrimeric combinations). Focusing on BAFF, a variety of hematopoietic and non-hematopoietic cell types produce/secrete BAFF, but – importantly – normal B cells do not. On the receptor side are TACI, BCMA and BAFF-R. According to the literature, all 3 receptors are expressed on B cells, although our own data in humans, including children, suggest that this is not the case (work in progress). The literature also indicates that BCMA is expressed on plasma cells, and TACI on macrophages and T-cell subsets. BAFF (and perhaps a truncated form of APRIL) bind to BAFF-R, while both BAFF and APRIL engage the other receptors.
| Varying BAFF Levels Post Rituximab |
|---|
 |
| Figure 2b: Varying BAFF results in 5 individuals post-rituximab (synchronized to the "1" time point). |
Our preliminary experience suggests that while there often are reciprocal relationships between B-cell counts and BAFF levels, this is not a set-in-stone rule. Several examples are shown to illustrate this further. In figure 2c (next page), two individuals are shown with different kinetics. The individual represented by the black squares received multiple doses of rituximab for autoimmune cytopenias on a background of DiGeorge syndrome. B-cell counts dropped precipitously, but have shown a quick and brisk recovery, in the context of sharply rising BAFF levels (that are still going up). The other individual received rituximab for an EBV-driven hemophagocytic lymphohistiocytic (HLH) disorder. In this case, Bcell counts exceeded baseline levels within months after multiple doses of rituximab, concurrently with only a modest increase in BAFF levels. This despite the fact that, as has been suggested in the literature, EBVinfected B cells can violate the "rule" that (normal) B cells do not produce BAFF (i.e. produce their own survival factor). As is clear from figures 2b and 2c, BAFF levels were already elevated at baseline. In fact, since the inception of measuring BAFF levels, it has been our experience that BAFF levels are often increased at baseline (especially in autoimmune cytopenias). This has been the experience in many studies in humans regarding BAFF levels in immunologic disorders (as nicely tabulated by Mackay et al, 2007).
| Different Profiles of BAFF and B-cell Recovery |
|---|
 |
The next two individuals (figure 2d, above), also show elevated pre-rituximab BAFF levels. In contrast to the previous examples, no significant further increase in levels was observed, despite the fact that B-cell depletion was complete and ongoing. Both subjects were treated for chronic ITP. Of note: these profiles of B-cell depletion without a subsequent rise in BAFF levels would suggest that the explanation, referenced in the literature, that BAFF levels sharply increase following rituximab, because of a sudden depletion of BAFF receptors, is not correct (or at least not complete).
It has also been suggested that high baseline BAFF levels would provide a set-up for rapid B-cell reconstitution and relapse of the autoimmune disorder. Our experience so far indicates that this is not the case: B-cell counts have not returned to normal in these and several other individuals with similar high baseline BAFF levels. In addition, no relapse of pre-rituximab conditions has been observed regardless of the speed of B-cell reconstitution and regardless of BAFF kinetics. How BAFF levels should be interpreted in combination with B-cell counts and B-cell phenotype remains far from straightforward. The conclusion that most safely can be drawn at this point is that many BAFF measurements, in combination with B-cell counts and other relevant measurements in large numbers of individuals with different disorders over a prolonged period of time are needed to become "un-baffled" by the BAFF system.
With every biomarker, one should consider the fact that variations in levels might be reflective of genetic variability that has the potential to influence the biology of the marker. Whether this also applies to BAFF or other constituents of the BAFF system will be a topic for a future discussion on genetic controllers of B-cell reconstitution.
