We investigated the role of the complement receptors CD21/35 in CWD prion accumulation, replication, and disease progression. We observed a complete rescue from terminal CWD of Tg5037 mice lacking CD21/35. Only 3 of 11 nonclinical Tg5037;CD21/35−/−
mice displayed detectable, yet reduced, prion neuropathology and PrPRES
deposition in their brains. These results reveal a more dramatic outcome than earlier studies showing only a partial rescue of CD21/35-deficient mice from scrapie infection, despite those mice expressing only WT (i.e., 5-fold less) PrPC
levels. This could reflect differences between mouse and cervid CD21 expression, as are apparent between mouse and human CD21. However, little is known about cervid CD21. The gene has yet to be cloned, so comparative analyses with murine CD21/35 are impossible at present. We can, however, compare CD21 sequence homology and phylogeny among other species that are susceptible to TSEs. For example, sheep, which are susceptible to scrapie, a TSE that closely resembles CWD in transmission efficiency, and lymphotropism, express a CD21 molecule that shares 65% sequence identity with murine CD21/35, including their ligand binding domains (). This may explain the similar lymphotropic characteristics of murine and ovine scrapie. Ovine CD21 also shares 65% identity with human CD21/35. Overall, CD21/35 from these three species share 52% identity and 64% similarity. In contrast, bovine CD21, which is 40% larger than the other three CD21/35 molecules (~1400 compared with ~1000 aa, respectively), shares <20% similarity to the other three CD21/35 molecules. Phylogenetic analysis reveals a clustering of murine, ovine, and human CD21/35 proteins, with bovine CD21 much more distantly related (). Interestingly, bovine spongiform encephalopathy has been shown to have little or no lymphotropic characteristics (44
), perhaps owing to the vastly different CD21 molecule that bovids express.
FIGURE 5. Sequence and phylogenetic analyses of CD21/35. (A) Sequence alignment of ligand binding domains of bovine, human (shown in lowercase letters), mouse, and sheep CD21. Dashes represent gaps in sequence alignment, revealing additional amino acids in bovine (more ...)
These results indicate a significant role in prion pathogenesis for CD21/35, the importance of which may vary by prion strain. Complement components C1q and C3 have recently been shown to exhibit similar strain preferences in vitro and in vivo (48
). We are currently investigating other prion strains to determine the contribution of CD21/35 to prion pathogenesis in those infection models. Interestingly, cross-species prion transmission was recently shown to result in a higher infection rate of the lymphoreticular system than the CNS in the xenohost (49
). This cross-species infection resulted in distinct lymphotropic and neurotropic strains with differential host ranges. These strains may result from tissue-specific strain selection or mutation. The higher efficiency of prion infection in the spleen (which harbors CD21/35-expressing FDCs and B cells) compared with the brain (which lacks them) alludes to a critical role for CD21/35 in prion retention, replication, and possibly strain selection in trans-species prion infection. The lack of CD21/35 that delays peripheral prion accumulation might further limit the lymphoid replication of neurotropic prion strains, resulting in delayed or abrogated disease progression. If so, this would have profound implications for prion xenotransmission and possible therapeutic approaches aimed at CD21/35. For example, targeting CD21/35 to slow the spread of neurotropic prions could be an attractive alternative to most prion disease therapeutics developed to date that target the CNS, which can complicate drug delivery. Interfering with CD21/35-mediated prion strain selection could also mitigate emergence of new prion strains with expanded host ranges and prevent a breach of the species barrier similar to the one that likely caused the bovine spongiform encephalopathy and subsequent new-variant Creutzfeldt-Jakob disease outbreak 15 y ago in the United Kingdom.
To study the kinetics of extraneural CWD prion accumulation, we amplified PrPRES
from spleens of CWD prion-infected Tg5037 and Tg5037;CD21/35−/−
mice at various time points throughout infection. At 15, 70, and 140 dpi and at terminal disease, prion accumulation was significantly lower in CD21/35-deficient mice. The extremely high prion load detected at 15 dpi most likely reflects increased retention of prion inocula early after infection. This delay in extraneural prion accumulation strongly correlates with abrogation of prion neuropathology and terminal disease. These results coincide with our previous data from scrapie mouse models (17
), further strengthening evidence that CD21/35 play an integral part in prion accumulation in peripheral lymphoid organs that ultimately facilitates neuroinvasion.
Furthermore, we show that CD21/35 are present in prion preparations enriched from spleen homogenates by NaPTA precipitation. We also demonstrate GC formation in spleens during prion infection primarily dependent on CD21/35 and PrPC
expression on FDCs. It may seem surprising that CD21/35 expression on FDCs, rather than B cells, correlates with prion-induced GC formation, because CD21/CD19/CD81 B cell coreceptor ligation helps activate B cells to form GCs. However, maximal B cell activation and GC formation require signaling from both the BCR and B cell coreceptor (32
). In this study, we show that although prion infection stimulates CD21/35 translocation to lipid rafts on B cells, signaling appears to be suboptimal for GC formation in the absence of concomitant BCR translocation. We observed a strong dependence on both PrP and CD21/35 expression on FDCs for a strong GC response. Paradoxically, CD21/35 translocation did not occur on FDCs, which are the major prion trappers and replicators but lack other B cell coreceptor components required for CD21/35 movement. One could therefore argue that GC formation represents an artifact, rather than being a driver of splenic prion replication. Elimination of GCs had no effect on peripheral prion replication and disease progression in mice infected i.p. with RML5 (50
), supporting this interpretation. However, GC-deficient mice infected intracranially with 139A mouse-adapted scrapie prions exhibited a significant delay to terminal disease (51
). Thus, distinct prion strains may differentially influence GC formation and subsequent prion pathogenesis. Additionally, this discrepancy further highlights potential preferences of distinct tissues for different prion strains. CD21/35-expressing cells within GCs may facilitate this selection process in the lymphoid system. Increased retention of prions on FDCs could induce a persistent state of prion presentation to adjacent B cells sufficient to cause an atypical GC response (40
). FDCs may coax B cells to linger there, providing increased lymphotoxin signaling to FDCs that may promote formation of hypertrophic dendrites that efficiently retain and replicate prions. Consistent with their role as long-lived, long-term APCs, FDCs may also present prions to neighboring PrPC
-expressing B cells that could induce CD21/35 translocation and move prions proximal to PrpC
in lipid rafts and promote further prion replication and GC formation.
Taken together, these data support a principal role of CD21/35 in peripheral prion pathogenesis by trapping PrPRES
on both B cells and FDCs. CD21/35 expression on FDCs remains of paramount importance in this process, with B cells playing a lesser, but still important, role. We have recently shown that few prion-bearing B cells transport prions from infection sites to draining lymph nodes, but their presence increased dramatically within lymph nodes, indicating a prominent role for B cells in intranodal prion trafficking (52
). We propose that CD21/35 mediate this and other crucial processes in lymph node prion trapping and replication and we are currently testing this hypothesis.