Comments for:

---

Re: Peripheral prion pursuit

Submitter: Laura Manuelidis | laura.manuelidis@yale.edu

Yale University

Published November 12, 2001

The dominant view of TSE agent invasion from the periphery into the brain is that B cells are essential. B cells have been proposed to act directly, by transporting the infectious agent, or indirectly, by promoting the development of some other critical cells such as follicular dendritic cells (FDC) and M cells. However, there are significant problems with this model since in many circumstances infection has been shown in the absence of FDC. Most notably, a year ago we showed lymphotoxin beta knockout (LTb ?/?) mice, which lack FDC and do not form the abnormal PrP that is otherwise seen on the FDC surface, nevertheless developed progressive CJD infections after intraperitoneal (ip) inoculation (J. Virol. 74:8614-22, 2000). These observations are consistent with our view that altered PrP represents a pathologic response to infection, such as that induced in a host receptor, rather than the infectious agent per se. Aguzzi did not mention these findings, but rather quoted his own group's work, which employed mice in which LTb was only incompletely inhibited. Partial LTb inhibition was reported to inhibit scrapie. However, this conclusion appears even more questionable since we have also shown positive neuroinvasion in additional relevant knockouts in which B cell, FDC, or M cell development are blocked (Proc. Natl. Acad. Sci. USA 98:9289-94, 2001, online in July).

What cells might carry these agents into the healthy brain? Several years ago we proposed macrophages and other cells of the myeloid lineage could carry CJD and other TSE agents from blood into the brain and other tissues, as they do in many other infections (Science 277: 94-98, 1997). Subsequently, we reported persistent CJD infection of myeloid (CD11b+) cells isolated from mouse spleens relatively early after ip inoculation (op cit., J. Virol. 2000). We believe that our data showing myeloid cell infectivity (macrophages and dendritic cells) are more definitive than those in the recent paper (JCI 108: 703-7, 2001) highlighted in the commentary of Aguzzi (JCI 108:661-2, 2001). Unlike the JCI studies of cells acutely isolated, we used 5 weeks of in vitro growth and propagation of myeloid cells to minimize non-specific contamination that accompanies tissue disruption.

Aguzzi also implies that the BSE agent in sheep may pose a unique and newly discovered transfusion problem. He quotes his own recent in vitro work on plasminogen bead binding of PrP fibrils and homogenate infectivity. However, he excludes prior relevant data on other agent strains. In 1978 our own laboratory, soon followed by, Tateishi's in Japan, Diringer's in Germany, and Gajdusek's in the USA, showed several CJD and scrapie agents can be retrieved from white blood cells of infected animals, even when the animals were not clinically ill. Furthermore, serum from infected animals shows considerably less infectivity than white cells, making the meaning of in vitro plasminogen binding less clear. Finally, myeloid cell development is flexible, and these cells respond uniquely to different environments. Thus, when myeloid cells in blood shuttle into target organs they may acquire different properties as residents in those tissues, e.g., as microglia in brain. Microglia can be filled with abnormal PrP (op cit., Science 1997), further supporting their infection. We have also detected abnormal PrP accumulation in dendritic and M cells, as identified by their morphology, in the gastrointestinal tract of Ig positive mice (op cit., Proc. Natl. Acad. Sci. 2001). Aguzzi states unequivocally that FDC are "stromal components" but respected immunologic experiments show FDC may also derive from hematopoietic cells. All these data tie infectivity to a variety of peripheral myeloid cell types. While emphasis continues to be placed on direct peripheral nerve contact for agent spread, we have recently shown that the CJD agent can travel hematogenously to the gut before it has a chance to replicate in the nervous system (BioMed Central Infectious Diseases 1:20, 2001).

Myeloid cell infection and tissue seeding has major health implications. These include 1) the multiple cellular reservoirs of agent that must be targeted in effective therapies, 2) the realization that, even with amplification techniques, abnormal prion protein may not be demonstrable in all infected cells (such as some white blood cells), and 3) cells of the gut infected very early may be shed into the environment with fecal matter, where they remain a source for further epidemic spread. The latter concern is most relevant for BSE since it is estimated that about 1 million cows have been infected.

It is very important for both the primary literature and commentary to reflect all of the work in a field, even if it appears to contradict a dominant view. In TSE research, there is a clear danger of moving away from this evenhanded approach, as the number of laboratories working with these infectious agent is very small, and the multinational collaboration/affiliations of the largest prion groups with each other can create a conflict of interest in the review process. If only the conventional views of this small number of groups are heard, a number of important implications and ideas, such as the role of myeloid cells in pathogenesis, could be missed. The consequences of this could be severe in the long run. Financial interests now involved in BSE may additionally divert work from fundamental issues to perhaps premature diagnostic or therapeutic claims based on manipulation of the prion protein in vitro, or on incompletely proven theories of spread in vivo.

---

Dendritic cells and prion pathogenesis

Submitter: Pierre Aucouturier | aucouturier@necker.fr

INSERM U25, Necker Hospital, Paris, France

Published November 5, 2001

The commentary article from Adriano Aguzzi (1), on our paper showing that dendritic cells allow infection of the central nervous system by prions (2), raises a number of interesting points.

After prions cross the gut epithelium, a "lympho-invasion" step allows these agents to amplify in lymphoid follicles, after which they can spread toward sites of "neuro-invasion". Dendritic cells are evident candidates for transporting prions at both the lympho- and neuro-invasion steps, a possibility that none of the numerous previous studies had even evoked.

It seems still too early to build definitive scenarios for prion neuro- invasion. First, there is a significant variety of prion strains that are thought to behave quite differently in cell targeting and pathogenesis. Second, even when considering a defined model of mouse scrapie, no single cellular or molecular entity has proved so far a full determinant for neuro-invasion. Thus, suppression of lymphotoxin-beta signalling (3), complement factors (4) or sympathetic innervation (5) only delays the onset of disease. microMT and RAG-/- mice, which are completely deficient in lymphoid follicles, do not get sick after peripheral inoculation with usual doses; however, in many tested animals prions do reach the brain (6), and much higher doses even provoke clinical scrapie (7).

What are then the actors of neuro-invasion? As mentioned by Adriano Aguzzi, RAG-/- mice display no evident qualitative or quantitative defect of DCs, and the above results from his own group are, contrary to his statement (1), strongly suggestive that the "endogenous pool of DCs" can actually support neuro-invasion by itself. Immature DCs are present in peripheral mucosa and are also numerous in secondary lymphoid organs. In normal mice infected with usual doses, accumulation inside lymphoid follicles would provide enough concentrated infectious material to be loaded by local DCs, while in RAG-/- mice much bigger inoculum would have to be given for mucosal DCs to perform neuro-invasion.

Finally, Adriano Aguzzi expresses concerns about the specificity of DC- triggered pathogenesis in RAG-/- recipients, noting that we could also infect animals using a B-cell enriched fraction. However, as we indicated, animals receiving the B-cell enriched fraction -- unlike the DC-injected recipients -- showed complete reconstitution of the lymphoid structures. Thus, neuro-invasion may not be solely related to the injected B cells.

1. Aguzzi, A. 2001. Peripheral prion pursuit. J. Clin. Invest. 108: 661-662.

2. Aucouturier, P., et al. 2001. Infected splenic dendritic cells are sufficient for prion transmission to the CNS in mouse scrapie. J. Clin. Invest. 108: 703-708.

3. Mabbott, N.A., Mackay, F., Minns, F., and Bruce, M.E. 2000. Temporary inactivation of follicular dendritic cells delays neuroinvasion of scrapie. Nat. Med. 6: 719-720.

4. Mabbott, N.A., Bruce, M.E., Botto, M., Walport, M.J., and Pepys, M.B. 2001. Temporary depletion of complement component C3 or genetic deficiency of C1q significantly delays onset of scrapie. Nat. Med. 7: 485-487.

5. Glatzel, M., Heppner, F.L., Albers, K.M., and Aguzzi, A. 2001. Sympathetic innervation of lymphoreticular organs is rate limiting for prion neuroinvasion. Neuron 31: 25-34.

6. Klein, M.A., Frigg, R., Flechsig, E., Raeber, A.J., Kalinke, U., Bluethmann, H., Bootz, F., Suter, M., Zinkernagel, R.M., and Aguzzi, A. 1997. A crucial role for B cells in neuroinvasive scrapie. Nature 390: 687 -690.

7. Klein, M.A., Frigg, R., Raeber, A.J., Flechsig, E., Hegyi, I., Zinkernagel, R.M., Weissmann, C., and Aguzzi, A. 1998. PrP expression in B lymphocytes is not required for prion neuroinvasion. Nat. Med. 4:1429- 1433.