Viromewide antibody responses after transplantation

In this issue of Blood, Bender Ignacio et al provide preliminary evidence for using a recently developed multiplex unbiased array (VirScan) to decipher humoral response after transplantation and the potential for VirScan to improve donor selection.¹ 

Transplant recipients are particularly susceptible to both viral reactivation and viral infection. After allogeneic hematopoietic stem cell transplantation (HSCT), immunosuppression used to prevent and treat graft-versus-host disease results in a heightened and prolonged risk for opportunistic viral infections²  (see figure panel A). Bender Ignacio et al (from the Fred Hutchinson Cancer Research Center [FHCRC]; Seattle, WA) report the first analysis of HSCT recipients who were sequentially followed with VirScan. The seminal article describing VirScan was published in Science by Xu et al³  in 2015. Basically, VirScan provides a comprehensive serologic profiling of human immunoglobulin G (IgG) to 206 viruses. This high-throughput technology allows detailed responses to viruses. It uses DNA microarray synthesis and bacteriophage display to create a representation of epitopes of the human virome. Immunoprecipitation and DNA sequencing are then used to characterize the peptides recognized as binding the IgG in the sample (see figure panel B). Since the original basic science article was published, the tool has slowly moved to translational research. Most recently, an article by Isnard et al⁴  described the temporal virus serologic profiling of kidney graft recipients using VirScan. In that study, which involved 45 kidney transplant recipients, serologic profiling was performed on day 0 and at 1 year. Results were compared with an enzyme-linked immunosorbent assay and a polymerase chain reaction assay. Antibody responses to 39 of 206 species of virus present in the library were detected, and these responses were largely conserved during the year after transplant, regardless of immunosuppressive therapy.

Bender Ignacio et al studied 37 patient-donor pairs sequentially through myeloablative transplant, including samples from pretransplant and at days 30, 100, and 365 posttransplant (see figure panel B, right portion). Donor age, donor-recipient cytomegalovirus (CMV) serostatus, and use of corticoids influenced the diversity of the IgG antibodies repertoire at day 100. Somewhat counterintuitively, the IgG repertoire was similar to that of the donor at day 100 but similar to that of the recipient at day 365. As expected, gain or loss of epitopes to common viruses differed by donor and recipient pretransplantation serostatus, with highest gains in naïve donors to seropositive recipients, in particular for herpesviruses and adenoviruses. As previously reported,⁵  CMV strongly shapes B-cell repertoire after allogeneic HSCT.

As always in good science, the Bender Ignacio article raises some questions. First, as a general comment, the authors used more sophisticated statistical analyses than those used in the kidney graft study.⁴  The statistical methodologies used in their study were developed for analyses of the microbiome, and so-called “ecologic metrics” were developed to describe not only the total epitope score but also the diversity within each individual (eg, Simpson’s D score that measures the α diversity), the donor-recipient antiviral response (β diversity), and the longitudinal estimate of distance between each donor-recipient pair using linear mixed-effects models, and to test the association between patient and transplant characteristics using generalized linear models. These refined biomathematical tools allowed the authors to perform a more nuanced analysis than that performed in the kidney transplant recipients (but much harder to read). As a cautionary note, these results are preliminary evidence because the number of patients studied is limited and the study involved only myeloablative conditioning. Moreover, the study is biased toward 1-year survivors and thus does not provide evidence about the IgG repertoire in patients who eventually succumbed as a result of viral-related diseases before 1 year.

What are the implications of those fascinating, although preliminary, results? The first implication is practical: VirScan may ultimately be a tool for screening and monitoring posttransplant virus infection. As stated above, it is too early to consider VirScan a routine method. In addition, as acknowledged by the authors, receiver operating characteristics of this synthetic virome to determine patient and donor serostatus is limited to viruses for which public epitopes and validated serologic methods are available (see supplemental Table 1 in the Bender Ignacio et al article for details). Furthermore, it should be remembered that VirScan analyzes only the IgG repertoire and does not investigate the Ig switch (Ig-M to Ig-G) that characterizes most recent viral infections.

The second perspective, in our opinion, is far more exciting. VirScan permits a greater in-depth analysis of humoral immune reconstitution after transplantation. B-cell reconstitution studies after HSCT have come of age in the past 10 years (reviewed in Sarantopoulos and Ritz⁶  and Socié⁷ ). Bender Ignacio et al address this point using only α and β metrics to correlate the IgG repertoire with 1-year total B-cell reconstitution (see supplemental Figure 4 in the article by Bender Ignacio et al for details). However, B-cell reconstitution is far more complex than could be ascertained from total B-cell counts (see figure panel C) (reviewed in Sarantopoulos and Ritz⁶  and Socié⁷ ). The early B-cell reconstitution is dominated by transitional B cells that are pregerminal center, nonswitched B cells. The naïve B-cell population (that does not secrete Ig) emerges only from 9 months to 1 year after transplantation, and it takes months (up to 2 years) to fully reconstitute memory B cells (and as an assumption, plasma cells, for which few if any immune reconstitution data are available).

Finally, the authors assumed that the average half-life of IgG from the recipient was 26 days and thus surmise that any significant level of virus-specific IgG should come from the donor. This assumption can be challenged because the allotype of the IgG has not been studied (although it is weakly polymorphic) after HSCT. Previous work by the FHCRC on hemagglutinin showed that recipient IgG can persist much longer than 1 month,⁸  and in 2007, a study demonstrated that humoral immunity to common viral and vaccine antigens can persist for decades (antibody half-life against rubella [114 years], Epstein-Barr virus [11.5 years], and varicella zoster virus [50 years]).⁹ 

In the near future, correlating in-depth cell phenotyping through mass cytometry and the B-cell receptor molecular rearrangements with results of the VirScan will be of major scientific interest.