A great deal of effort has gone into studying how peripherally acquired prion diseases make their way into the brain [reviewed in Mabbott & Macpherson 2006]. In general, it is thought that prions acquired peripherally – whether orally, subcutaneously or intraperitoneally – make their way to the lymphatic system as a staging point. They colonize follicular dendritic cells, replicate in the spleen, tonsils, lymph nodes, and appendix, and after a great deal of time eventually infect the peripheral nervous system and find their way into the central nervous system. There is evidence for other routes as well – for instance, neuroinvasion from the tongue bypassing the lymphoreticular system [Bartz 2005].
In contrast, it is not quite as clear where infectivity originates in spontaneous prion diseases – by which I mean, sporadic and genetic prion diseases.
obviously, it starts in the brain
Perhaps the most obvious reason to assume that infectivity starts in the brain is that PrP is expressed more highly in the brain than in other tissues. See for instance the Human BodyMap 2.0 RNA-seq data, the PRNP gene card (though one of the methods there reports higher expression in skin) or Novartis BioGPS. Since prion replication depends on PrP expression level, it seems parsimonious to figure that spontaneous infectivity would originate in the tissue with the highest expression level.
But if we’re not happy with assumptions (and we shouldn’t be), another source of evidence we can look to is studies of infectivity in peripheral tissues in spontaneous prion disease. The results in this area are mixed, largely because our ability to detect infectivity has improved so dramatically over the years.
In the bad old days before we knew what prions were, the only way to test infectivity was to infect an animal. The earliest experiments injecting chimpanzees and other non-human primates with homogenates of various human tissues didn’t find very much infectivity in any peripheral tissues, even in kuru, which of course was a peripherally acquired prion disease [Table 10 in Brown 1994].
With the advent of the BSE / vCJD epidemic and the threat of transmission in the U.K. population, there started to be a great deal of interest in whether prions were in peripheral tissues. Studies found loads of PK-resistant PrPSc in spleens, tonsils and lymph nodes of variant CJD patients – who had acquired prion disease peripherally – but no PK-resistant PrPSc at all in the same tissues in sporadic CJD patients [Hill 1999, Head 2004]. That finding is quite consistent with the idea that spontaneous prion disease starts in the brain and stays in the brain. Peripheral nerves in genetic prion disease patients (CJD, GSS and FFI) likewise appeared devoide of PrPSc [Hainfeller & Budka 1999]. Only the eye appeared to contain PrPSc in sCJD [Head 2003].
With the discovery that NaPTA could concentrate PrPSc [Safar 1998] it became possible to detect PrPSc much more sensitively in a Western blot. Using NaPTA, it was discovered that there was in fact some amount of PrPSc in the spleen and skeletal muscle in some, but not all, sporadic CJD patients [Glatzel 2003]. Interestingly, the patients with PrPSc in peripheral tissues in their peripheral tissues by the time of death were those who had had a longer disease duration. (While the sample size was small, it also appeared that patients with any sCJD subtype other than MM1 were more likely to have peripheral PrPSc.) Using similar methods, another study around the same time found PrPSc in the olfactory epithelium [Zanusso 2003], the place where olfactory nerve fibers from the brain interface directly with the outside world.
So while sporadic CJD appeared to lack the huge accumulation of prions in the spleen seen in vCJD, there was indeed some degree of peripheral prion infection going on in sporadic cases. But it was mostly in the tissues most closely connected to the brain – the eyes, nose, and so on. This seemed more likely an effect, rather than a cause, of the brain infection. One author argued that the lymphoreticular involvement in vCJD was evidence for centripetal spread into the brain, while the limited peripheral involvement in sCJD was best interpreted as centrifugal spread – prions coming out from the brain to infect those tissues most closely associated with the brain [Head 2004].
One way to confirm this theory would be to see if intracerebral inoculation – where you know for sure the disease starts in the brain – produces the same pattern of peripheral infection as seen in sCJD. Someone has probably done precisely that experiment, though I didn’t find it in my brief, non-exhaustive literature search. One primate study toyed with this question, but confusingly, the animals were inoculated both intracerebrally and in the tonsils [Herzog 2005]. PMCA provided an even newer, more sensitive way to detect PrPSc, and a study on peripherally infected hamsters found that infectivity in the blood rose initially, dropped off as the infection progressed, and then spiked up again at the onset of symptoms [Saa 2006]. How the original peripheral infectivity had been cleared was never addressed, but the fact that infectivity rose again at the onset of symptoms was attributed to “brain leakage”.
does that settle it?
While the studies reviewed above are not 100% conclusive, it certainly sounds quite plausible that the peripheral PrPSc observed in sCJD and genetic prion diseases is, as Head put it, “centrifugal,” arising from the brain rather than being left behind by prions on their way to the brain. Particularly if you think, as I do, that origin in the brain is the most likely scenario a priori, then the data are pretty consistent with that.
Yet I’m a bit disturbed that this has never really been conclusively shown in the literature. If we really wanted to be sure, it’s an easy enough experiment to do. We now have several mouse models of spontaneous prion disease [Hsiao 1994, Jackson 2009, Yang 2009, Friedman-Levi 2011, Watts 2012] and we also have treatments that are capable of dramatically slowing or even completely curing peripheral prion infection (e.g. pentosan polysulfate [Ehlers & Diringer 1984] and antibodies [White 2003] respectively) despite not crossing the blood-brain barrier and not being effective if peripherally administered against CNS prion infections [Doh-Ura 2004, White 2003]. So: treat those mice with those treatments and if disease is delayed, then the infectivity must have been arising outside the brain.
I’d be willing to bet pretty long odds that the answer is yes, it starts in the brain, and no, those treatments don’t work. But on the off-chance that we’re all wrong, that would be kind of a big deal.