<<

. 37
( 45 .)



>>

Again, although statisttcally significant changes m scrapie mcubation time
can be observed when transferrmg prtons between nomsogentc mouse strams,
these effects have an explanation that lies outside of the realm of transplanta-
tion biology (Z2). The factors that do control prion host-range will be consid-
ered m the followmg, after a description of the origins and defimtion of “species
barrier” effects.
2. Species Barriers to Prion Transmission
Although first described m rodents, “species barrier” effects have been con-
firmed m many expertmental settmgs (13). The general observatton is that the
first transmission of prtons from one species to another results in incubation
times that are more protracted than subsequent repassagesm the recipient spe-
cies* Adaptation to the new species frequently IS complete by the second pas-
sage (1446). There are variations on this theme: Some species barriers are
very hard to traverse, wtth mcubatton times approaching the natural lifespan of
the recipient, often with only a fraction of the inoculated animals exhibiting
disease, e.g., transmission of CJD prtons to mice (17,18) or hamsters (19).
Some barriers appear msurmountable, for example, passage of the hamster-
adapted scrapie priori isolates 263K and Sc237, or transmtsstble mink
encephalopathy prions into mice (2&22).
In addition to laboratory studies, species barriers operating in natural prion
diseases can be tentatively inferred from epidemiologtcal data. For example,
even though natural scrapie 1sendemic m several European countries and the
United States, handling of or consumptton of meat from affected sheep does
not appear to be a significant risk factor for sporadic CJD (23,24). CJD cases
have yet to be reported in laboratory workers handling rodent-adapted scrapie
isolates but have been described m histopathology technictans handling human
biopsy material (25).
3. PrP Sequences and Species Barriers
The hypothesis that primary structure of PrP plays a pivotal role in species
barrier effects was derived from several pieces of information that had become
Pi-ion “Species-Bamer” Effects 253

available by the late 1980s. These included the observations that PrP is a maJor
component of purified preparations of scrapie prions (26) and that different
scrapie-susceptible mammals encoded nonidentical PrPs (27-29). It was also
known that scrapie mcubation time gene alleles m mice were correlated with
two alleles of the prton protein gene, Prnp: These a and b alleles diverge at
codons 108 (Leu + Phe) and 189 (Thr + Val) (30). Furthermore, isolates of
prions derived from Prnp“ mice differed in duration and variance of scrapie
incubation period (when repassaged in Prnpa mice) from those derived from
Prnph mice (12). Taken together, these data suggested that optimal replication
of prions might be dependent on “matching” incoming PrPSC with endogenous
PrPC such that both types of molecules shared the same primary structure.
Although transgemc studies that directly address this hypothesis will be scruti-
nized presently, recent compilations of PrP sequences are compatible with this
interpretation.
Transmission of CJD prtons into nonhuman primates is correlated with the
presence of Met at codon 112 (31). Data from a smaller study is in accord with
this finding, even though these authors reached the opposite conclusion that
“primary structural similarity may not be an important factor m disease trans-
mission”: This divergent interpretation may reflect considerations detailed m
Section 7., absence of Capuchin monkey and marmoset PrP sequences from
the compilation, or presence of PrP-coding polymorphisms m the particular
Rhesus monkeys used for maculations, which exhibited a transmission rate of
only 73%, yet (in other individuals of thts species) 96 4% ammo-acid homol-
ogy to human PrP (32). Other studies have elaborated on the concept that PrP
gene alleles within a species can comprise an “mternal” (mtra-) species barrier
to prion mfectton. Coflinge and coworkers have demonstrated that Val/Met
heterozygosity at codon 129 in the human PrP gene may confer a degree of
resistance to the development of CJD (33). Similarly, a common Gln/Arg poly-
morphism at codon 171 m the sheep PrP gene may comprise a major determi-
nant of susceptibility to natural scrapie in Suffolk sheep (34).
4. Studies with Syrian Hamster PrP Transgenes
To address the relationship between PrP primary structure and species bar-
rier phenomena, a cosmid clone encompassing a complete Syrian hamster
(SHa) PrP gene was used to construct transgenic mice (3.5). Four transgemc
lmes were shown to express SHa PrP mRNA and PrPC: All exhibited height-
ened susceptibihty to infection with hamster-adapted prions. Furthermore,
incubation times were inversely related to the level of SHaPrPC expression: In
the cases of the Tg(SHaPrP)81 and Tg(SHaPrP)7 lines these times equalled
(75 d) and surpassed (55 d), respectively, those observed in hamsters. Titers of
mouse prions m Tg mice inoculated with SHa-adapted prions were low or
254 Westaway

undetectable Conversely, SHa prions were undetectable in the same Tg mice
when inoculated wtth mouse-adapted prtons. Patterns of spongtform change
and amylord deposition in TgSHaPrP mice inoculated with SHa-adapted prrons
closely resembled those of hamsters (2Z). These results indicated that a highly
specific interaction between a transgene encoded molecule (presumably PrPC)
and SHaPrPSC the moculum features in prion rephcatlon. Thrs hypothesis
m
was appraised in a further set of experiments.
By engmeermg restrlctron sites mto mouse (MO) and SHa PrP cDNA clones,
a “mix-and-match” approach was used to create chtmertc PrP coding cassettes.
These were then inserted into a PrP expression vector, cosSHa Tet, where the
coding sequences of the SHa PrP cosmid have been replaced by a tetracycline
resistance gene flanked by SalI restriction sites (36). One such construct har-
boring the MH2M cassette,which differs from mouse PrP at residues 108, 111,
138, 154, and 169 (i.e., m the central region of the PrPCmolecule), engendered
the Tg(MH2M)92 transgemc hne. Interestmgly, this lme was found to be sus-
ceptible to both SHa- and mouse-adapted prtons. Furthermore, prion isolates
that transited Tg(MH2M)92 mice gained novel host-range properties. For
example, subsequent to passagem Tg(MH2M)92 mice, the Sc237 SHa-adapted
isolate was able to infect mice with an incubation period of approx 185 d.
Conversely, the mouse-adapted Rocky Mountain Laboratory prlon isolate
(RML) gamed the ability to infect Syrian hamsters with an mcubatron period
of approx 215 d (37).
Transgemc lmes expressing a chtmertc PrPC with a smaller hamster-derived
coding region (differences only at codons 108 and 111) were not susceptible to
Sc237 prtons. These experiments demonstrate the crucial importance of cer-
tam ammo acid side chains in the central region of the PrP m determining host-
range and further implicate homophilic PrPsc/PrPc mteracttons in prton
replication. Addmonally, since the only differences between the constructs m
Tg lmes, such as Tg(MH2M)92 and Tg(MHM2)285, he m the PrP coding
region, these expertments dispense with an earlier caveat that other genes or
sequenceswtthm the SHaPrP cosmtd control drssolutton of the barrier to infec-
tion with Sc237 prions (35).
5. Studies with Human and Sheep PrP Transgenic Mice
Extrapolating from the precedent established by Tg(SHaPrP) mice, attempts
have been made to transmit human and sheep prtons to cognate TgPrP mice.
Transgemc lines Tg(HuPrP)l 10 and Tg(HuPrP)l52 were constructed by
inserting wild-type human PrP coding sequences encoding Val at codon 129
into cosSHa Tet. For the Tg(HuPrP)152 line levels of human PrPC were
approximately four- to eightfold higher than that found in human brain. These
Tg mice were challenged with mocula from casesof Gerstmann-Straussler syn-
Prion “Species-Barrier” Effects 255

drome (GSS), as well as sporadic and iatrogemc CJD. With the exception of
transmission from iatrogemc CJD case #170 into Tg(HuPrP)l52 (positive at
approx 240 d, J. Collmge, personal communication), mcubation times were
protracted (2589 d+ SE), with only 14/169 inoculated mice exhibiting scrapie-
like clmical symptoms. Notably, a simrlar (10.3%) disease incidence was
encountered m non-Tg controls (18).
Difficulties have also been encountered transmittmg sheep prions to Tg(PrP)
mice. Thus, the Tg(ShePrP)217 lme expressing wild-type sheep PrPC (corre-
sponding to the Gln 171 allele) did not exhibit heightened susceptibility to
natural scrapie prions from Suffolk sheep, smce both transgenic and nontrans-
gemc ammals became sick at 2360 d after inoculation (34,38,39), (D
Westaway, D. Foster, and S. B. Prusmer, in preparation). Thus, experiences
with Tg(HuPrP) and Tg(ShePrP) mice suggest that expression of a donor-
derived PrP transgene is not always sufficient to erase species barriers to prion
infection.
6. Enhancing Transmission of Exogenous Prions into TgPrP Mice
One explanation for the “failures” with human and sheep transgenes detailed
above focuses on the contribution of endogenous PrP genes. Incubation times
to illness in TgSHaPrP lines Inoculated with RML mouse prrons range from
18-38 (Tg69) to 4666 (Tg8 1) d longer than m non-Tg littermates. This effect
crudely parallels levels of transgene expression, suggesting that the presence
of SHaPrPC can interfere with the conversion of MO PrPC to MO PrPSC (21).
This explanation is further supported by:
1 Approximately 40 d prolongations of incubatron times with RML prions in
Tg(MHu2M)PrP mice compared to non-Tg mice (see the followzng [18]),
2. Studies m scraple-Infected mouse neuroblastoma cells, where expression of heter-
ologousPrPsequences attenuate
can production of protease-resistant (36); and
PrP
3. Deletion of endogenous PrP coding sequences

Thus, mtroduction of the Tg81 SHaPrP transgene array onto a Prn-p”/o
ablated background led to successive reductions m the time from maculation
with Sc237 prions to onset of clinical disease from 75 f 1 d (Prnp”˜) to 67 + 4 d
(Prn˜+˜˜) to 56 + 3 d (Prn˜˜˜˜) (40,41). Analogous experiments to assessthe
behavior of Tg(HuPrP)l52 and Tg(ShePrP)217 transgenes within a PrnpO™O
background are in progress
Since transgene expression levels have a potent effect on mcubation times
(22,42), another maneuver is to construct Tg lines with yet higher levels of
expression. Unfortunately, the highest levels of PrPC expression obtained to
date are associated with a spontaneous neuromuscular disease. In the case of
Tg(SHaPrP)7+˜- heterozygous mice this syndrome has an onset after 700 d, and
256 Westaway
does not affect the execution or interpretation of experiments where animals
are inoculated with Sc237 SHa-adapted prtons. However, there 1sa more tan-
gible problem m the case of Tg(ShePrP) mice. Although several Tg(ShePrP)
lines equal the levels of ShePrP mRNA seen m Tg(ShePrP)2 17 mice, founder
mice wtth higher expression levels and copy numbers are rare (n = 2) and
engender unstable lines that “revert” to a Tg(ShePrP)2 17-like copy number
(39). Thus, the caveats of spontaneous disease and genetic instability assoct-
ated wtth high copy-number transgene arrays-with PrPc-assoctated cytotox-
tctty perhaps acting as the selection pressure for the emergence of lower
copy-number revertants-hmtt the practicality of this approach.
A thud enhancement to transgemc strategies mvolves use of chimertc cod-
mg sequences. Insertion of the central region of human PrP mto a mouse cod-
mg cassette to create a “MHu2M” transgemc line resulted m mcubatton times
to illness of 1240 d subsequent to maculation with CJD prtons (18). All the
maculated mice developed dtsease (n = 24) and mcubatton ttmes in mdtvidual
experiments were clustered (standard error of the mean f 4.6 d or less), mdica-
tive of a nonstochastic process. Demonstration of spongtform degeneration
typical of priori diseasesand accumulatton of transgene-encoded PrPSC detected
by a monoclonal antiserum argues that the Tg(MHu2M)PrP mice are more
susceptible to CJD prtons than their nontransgemc relatives.
7. Beyond Primary Structure:
Other Components of Species Barriers
Although the usefulness of Tg mice susceptible to heterologous prtons needs
no elaboration, tt should be clearly understood that parameters other than PrP
primary structure can affect transit across a species barrier.
7.1. “Strain “-Type
Different prton Isolates or “strains” propagated wtthm the same inbred host
may have different abilnies to cross a species barrier, e.g., hamster-adapted
scrapie strains 263K and 43 1K (20). Ortgmally attributed to the presence of a
nucleic acid genome, the molecular basis of “strainness” is unresolved but in
the case of scrapte and TME “strams” is associated with PrPSC molecules with
dtstmct phystochemtcal properties (43,44).
7.2. Chaperones/Foldases
Failure of Tg(HuPrP) mice to succumb to CJD prtons has been Interpreted
to reflect divergence m a gene encoding a chaperone or foldase involved in the
conversion of PrPC to PrPSC (18). Although aspects of induction of the heat-
shock proteins (hsps) hsp72, hsp28, and hsp73 are now known to differ between
prton-infected and uninfected neuroblastoma cells (4.5), these hsps seem
Priori “Species-Barrier” Effects 257

unltkely candidates for PrP-specific chaperones: They are not known to have
affinities for particular mature (completely folded) proteins, instead having a
catholic affinity for denatured proteins (46). Nonetheless, because there is at
least one example of a “dedicated” foldase in a eukaryote, a putative rhodopsin-
specific prolme tsomerase from D melanogaster (47), this lme of reasoning
may warrant further attention.
7.3. Carbohydrate Modifications
PrPSC known to be decoratedwith two complex Asn-linked carbohydrate trees
is
(48) and contams a glycan core wtthm the phosphattdylmostol glycohptd anchor
(3). Smcemammalian speciesdiffer m their repertoire of glycosyl transferases(49)
tt is possible that species-spectfic patterns of glycosylatton of PrPSC affect
also
the abthty to infect a new host. Indeed, the related notion that priori strains could
correspond to different PrPSC glycoforms 1squite widely held (43,50,51).
8. Reprise: General Approaches to Creating Tg(PrP) Mice
Experiences set out herem can be distilled into the following recommenda-
tions. Most or all of these pomts should be considered before embarking on
transgemc studies.
1 Probes Nucleic actd and antibody probes directed against the heterologous PrP
gene should be available. Ideally, these should exhibit no crossreactivity with
mouse PrP mRNA or PrPC In the case of cosTet SHaPrP the diverged 3™
untranslated region has proven a useful probe for DNA and RNA analyses (36)
2 The transgene construct Because expression of “mimgene” transgene constructs
IS weak or unpredictable (3.5,52), molecular clones encompassmg mtrons and
many kilobases of flanking sequence are recommended A variety of PrP cosmids
has been used to date. These have important advantages m that most stable Tg
lines express PrPC, 1.e , expression of the cosmids appears independent of chro-
mosomal context (21,39,42), and since they use PrP promoters, they direct
expression of PrPC to the correct cellular lineages. Procedures for the production
of transgemc mice are detailed by Hogan et al (53)
3. Allele-matching. If the species under study exhibits PrP coding sequence poly-
morphisms, care should be taken to derive the transgene construct from the
appropriate allele of the donor species
4 Genetic background. Use of PrnpO™O(Prnp “ablated”) mice is recommended (40)
This gene knockout was origmally created on the 129lSv background but because
this strain has low fecundity and oocytes that are difficult to micromlect, repeated
backcrossmg to transfer the ablated Prnp locus to PrP Tg mice within the FVB/N
or C57B16 backgrounds should be considered.
5 Expression levels. In general, stable Tg lines with high levels of expression (24 x
endogenous PrPC) are recommended for btoassays, because they exhibit the most
rapid onset of clmtcal disease Unmoculated Tg mice should be set aside to mom-
258 Westaway
tor for the appearance of spontaneous (as well as intercurrent) dtsease (39) Breed-
mg homozygotes for the transgene array offers the posstblhty of obtamrng yet
hrgher expression levels but there are attendant comphcattons
a Since transgenes can be mserttonal mutagens there IS a posslbthty of creatmg
homozygotes for n-relevant recessive lethal mutations (.54),
b DNA or segregation analyses ˜111 have to be performed to dtstmgutsh het-
erozygotes from homozygotes,
c There is the posslbtltty of creating a spontaneous PrP transgene-mediated
neurodegenerattve disease (see earlier), and
d Unequal crossmg over may destabilize the transgene arrays (which usually
correspond to tandem repeat units [54/)
6 Chtmertc or nonchtmertc PrP coding sequences7 Tg(MHu2M)PrP mice have
estabhshed the utility of chtmertc coding sequences Such mace mtght be constd-
ered for use m a sensitive bioassay for CJD prtons but use m other contexts mtght
be mapproprtate For example, the btophystcal properttes of “arttfictal” prlons
containing MHu2M PrPCJD may differ from those of authentic CJD prtons
7 Disease dtagnosts Procedures for moculattons and the chmcal dtagnosts of scrapte-
hke disease m rodents have been described elsewhere (12,16,18,55). “Posmve” dtag-
noses typtcally should be verified by htstopathology and the presence of PrPSC Mouse
colonies should be assessed for and protected against extraneous pathogens (53)
8 Biohazards Last, but not least, prtons cause fatal neurologtc disease m humans
and animals and should be handled with due respect Safety Issues are considered
by Taylor (Chapter 6) Laboratory studtes should be undertaken under the aus-
pices of local biohazard and recombmant DNA advisory commtttees and readers
are also referred to refs 56 and 57

9. Concluding Remarks

<<

. 37
( 45 .)



>>