It is generally thought that effective treatments for prion diseases need to inhibit prion propagation, protect neuronal tissues and promote functional recovery of degenerated nerve tissues. In addition, such treatments should be effective even when given after clinical onset of the disease and administered via a peripheral route. In this study, the effect of peripheral administration of an anti-PrP antibody on disease progression in prion-infected mice was examined. mAb 31C6 was administered via the tail veins of prion-infected mice at the time of clinical onset (120 days post-inoculation with the Chandler prion strain) and the distribution of this mAb in the brain and its effect on mouse survival assessed. The antibody was distributed to the cerebellums and thalami of the infected mice and more than half these mice survived longer than mice that had been given a negative control mAb. The level of PrPSc in the mAb 31C6-treated mice was lower than that in mice treated with the negative control mAb and progression of neuropathological lesions in the cerebellum, where the mAb 31C6 was well distributed, appeared to be mitigated. These results suggest that administration of an anti-PrP mAb through a peripheral route is a candidate for the treatment of prion diseases.
glial fibrillary acidic protein
ionized calcium-binding adaptor molecule 1
cellular prion protein
fusion protein between PrPC and the Fc portion of immunoglobulin
disease-specific isoform of the prion protein; scFv, single-chain fragment variable antibody
Prion diseases are fatal neurodegenerative disorders of humans and animals that are strongly associated with conversion of PrPC to PrPSc. Based on in vitro and in vivo studies, researchers have proposed many inhibitors of PrPSc formation, including anti-PrP antibodies, as potential therapeutic candidates . However, there are still no effective treatments for prion diseases.
Anti-PrP antibodies inhibit prion propagation in cells that are persistently infected with these agents [2-7]. Researchers have also demonstrated an inhibitory effect of anti-PrP antibodies has in vivo: provided antibodies are administered shortly after prion inoculation, intraperitoneal administration of anti-PrP antibodies prevents prion infection via a peripheral route but is not effective against intracerebral prion inoculation . Moreover, active immunization with recombinant PrP, a synthetic PrP peptide or a DNA vaccine following peripheral prion infection reportedly delays onset of the disease in mice and is a prerequisite for obtaining a prophylactic effect [9-12]. These results suggest that anti-PrP antibodies can prevent establishment of prion infection in peripheral tissues and thus may be useful for post-exposure prophylaxis.
We recently showed that, even when initiated after clinical onset, intraventricular infusion of an anti-PrP mAb slows disease progression; however, the effect of the mAb differs with different prion strains . In the intermediate stage after intracerebral prion inoculation, expression of a PrP-Fc or scFv in brains of mice that have been infected with prions by a using viral vector also inhibits disease progression [14, 15]. Taken together, these results indicate that immunotherapy is a potential candidate for the treatment of prion diseases, provided the antibodies can be efficiently delivered into the brain parenchyma.
In practice, peripheral administration of antibodies is desirable because it prevents or reduces the risk of prion contamination and dissemination by neurosurgical procedures. However, there have been no reports that peripheral administration of anti-PrP antibodies initiated at a late stage of infection is effective in preventing prion propagation and disease progression. Therefore, we administered anti-PrP mAbs through the tail veins at 120 dpi, after the time of disease onset, to determine whether peripheral administration inhibits disease progression.
MATERIALS AND METHODS
The anti-PrP mAb 31C6 (IgG1), which recognizes amino acids 143–149 of mouse PrP , and the isotype-matched control mAb P2-284 (IgG1) against feline parvovirus  were used. Purification of the mAbs from the supernatant of a serum-free high-density hybridoma culture was carried out as described previously . Labeling of the purified mAbs with a fluorescent dye was carried out using an Alexa Fluor 488 Protein Labeling kit (Molecular Probes, Eugene, OR, USA). The following rabbit polyclonal antibodies were used as primary antibodies for immunohistochemistry: B103 for detection of PrPSc [13, 18], anti-glial fibrillary acidic protein (GFAP; Dako, Carpinteria, CA, USA) and anti-ionized calcium binding adaptor molecule 1 (Iba1; Wako, Osaka, Japan) for detection of astrocytes and microglia, respectively.
Mice and prion strains
Protocols for the animal experiments were approved by the Institutional Committee for Animal Experiments. The mouse-adapted scrapie Chandler strain was used in this study. For intracerebral inoculation, 4-week-old female Jcl:ICR mice were purchased from CLEA (Osaka, Japan. Twenty microliters of 10% brain homogenate from mice infected with the Chandler strain were injected into the left cerebral hemispheres of Jcl:ICR mice after 2 weeks acclimation.
Antibody solution (2 mg/mL in PBS) was injected into the tail veins (0.25 mL/mouse) every week from 120 dpi. The mice were weighed at each time of administration. When they had lapsed into the terminal stage of the disease, or in some cases at 145 dpi, the mice were killed under anesthesia and their left and right brain cerebral hemispheres subjected to immunoblotting and histopathological examination, respectively.
Histopathology and immunohistochemistry
Embedding, sectioning, HE staining, and immunohistochemistry were carried out as described previously . Quantitative analyses were performed using Image J (Rasband, W.S. Image J, U. S. National Institutes of Health, Bethesda, MD, USA, http://rsb.info.nih.gov/ij/, 1997–2009). GFAP positive cells (astrocytes) and Iba-1 positive cells (microglia) were counted in stained areas over 2 × 10−5 mm2 in size. Vacuoles bigger than 2 × 10−5 mm2 were also measured to assess vacuolation.
Mice administered Alexa Fluor 488-labeled antibodies via a tail vein at 120 or 140 dpi were killed 3 days after antibody administration. Their brains were embedded with Tissue-Tek OCT compound (Sakura, Tokyo, Japan) and frozen at −85°C. Cryosections (10 µm thick) were prepared and were fixed with acetone for 10 min. Sections were mounted with Vectashield containing propidium iodine (Vector Laboratories, Burlingame, CA, USA) and were examined using a C1 confocal laser scanning microscope (Nikon, Tokyo, Japan).
Detection of total PrP and PrPSc as proteinase K-resistant PrP by immunoblotting was carried out as described elsewhere [13, 19].
Distribution of anti- prion protein monoclonal antibodies in mouse brains after intravenous administration
Jcl:ICR mice inoculated with the Chandler strain show early clinical manifestations of prion disease, such as ataxia of the hind limbs and changes in pelage and posture at around 120 dpi and lapse into the terminal stage of the disease at 150 dpi. To evaluate the therapeutic potential of anti-PrP mAb in the late stage of this infection, we started mAb administration via tail veins at 120 dpi and subsequently assessed PrPSc accumulation and neurohistopathological lesions.
We first examined the distribution of Alexa Fluor 488-labeled mAb 31C6 administered via tail veins. We detected fluorescence signals in the cerebellar medullas of mice infected with the Chandler strain 3 days after mAb administration when antibody was administered at 120 dpi. When administered at 140 dpi, the signals were more intense in the cerebellar medullas and we also detected them in the thalami. In addition, we detected weak fluorescent signals in the hippocampus (Fig. 1). In contrast, we detected no fluorescent signals in the brains of age-matched uninfected mice that had received Alexa Fluor 488-labeled mAb 31C6 (Fig. 1, Mock-infected, 140 dpi). These results demonstrate that mAb 31C6 can reach the brain parenchyma when administered via the tail veins after clinical onset of prion infection in mice.
Effect of intravenous antibody administration on mouse survival
Figure 2 shows the survival curves of Chandler strain-infected mice i.v. administered mAb 31C6 (anti-PrP mAb) or P2-284 (negative control mAb). The mean survival time of Chandler strain-infected mice that had received mAb 31C6 via a tail vein starting from 120 dpi (158.3 ± 9.6 days, n = 12) was longer than that of those that had received negative control mAb (153.3 ± 4.3 days, n = 9). Although the difference in survival of these groups was not statistically significant, it is of note that 7 of the 12 mice that had been given mAb 31C6 survived longer than 153 dpi, which was the mean survival time of mice administered the negative control mAb. The mean survival time of these seven mice was 165.4 ± 5.4 days.
Loss of body weight with disease progression is a prominent feature in Chandler strain-infected mice. To evaluate more objectively the effect of the anti-PrP mAb on disease progression, we weighed the mice every week after 120 dpi. At around 140 dpi, the decreasing gradient in body weight of the seven mAb 31C6-treated mice that had survived longer than 153 dpi was less steep than that of the negative control mice. In addition, they weighed significantly more than the negative control mice at 148 and 155 dpi (Fig. 2b).
We observed no prolongation in survival in 5 of the 12 mice that had been given mAb 31C6 (Fig. 2a). To investigate the reason for this, we analyzed the distribution of mAb in each mouse brain using immunofluorescence techniques. We detected no mAb in the brains of mice that had received negative control mAb (Fig. 3, P2-284). Although we detected mAb 31C6 in the cerebellums and thalami of mice that had no prolongation of survival time (Fig. 3, 148 and 149 dpi), the intensity and range of these fluorescence signals were much less than in mice that survived for longer than 153 dpi (Fig. 3, 167 and 172 dpi). These results suggest that i.v. administration of anti-PrP mAb could slow down disease progression in mice in which mAb is well-distributed through the brain parenchyma.
Effect of antibody administration on accumulation of the disease-specific isoform of the prion protein and neuropathological lesions
To analyze whether mAb 31C6 administered through the tail veins inhibits PrPSc formation in the brain, we initiated i.v. administration of mAb 31C6 at 120 dpi and used immunoblot analysis to quantify total PrP and proteinase K-resistant PrP at 145 dpi and in the terminal stage of the disease. As shown in Figure 2, some of the mice given mAb 31C6 did not survive longer than those that had been given negative control antibody; this finding was not unexpected because mAb 31C6 was inefficiently distributed through the brain parenchyma of these mice. Therefore, we first evaluated the effect of mAb 31C6 by changes in body weight, which parallel disease progression (Fig. 4a). Compared with the negative control group (given mAb P2-284), the experimental group (given mAb 31C6) had lost weight more slowly by around 141 dpi. The difference at 141 dpi was significant, indicating that the anti-prion effect of mAb 31C6 was evident in all mice in this experimental group.
By 145 dpi, the mean level of PrPSc in the negative control mAb-treated group had increased 2.4-fold over that at 120 dpi, whereas the mean level of PrPSc in the mAb 31C6-treated group (1.8-fold increase) appeared to be less than that of the negative control at the same time point (Fig. 4b, c). However, in the terminal stage of the disease, we found no differences in mean level of PrPSc between P2-284- and 31C6-treated groups. These findings suggest that mAb 31C6 slows accumulation of PrPSc in the brains of prion-infected mice. In contrast to the changes in the mean level of PrPSc, the mean total PrP of the mAb 31C6-treated group was significantly greater than that in the mAb P2-284-treated group. Compared with the level at 120 dpi, the mean total PrP in the mAb 31C6-treated group had increased 2.0-fold at 145 dpi and 2.3-fold by the terminal stage of the disease, whereas that of P2-284-treated group had only increased 1.4-fold at 145 dpi and 1.5-fold at the terminal stage (Fig. 4b, d).
Although immunoblot analysis showed that the mean level of PrPSc in the mAb 31C6-treated group was approximately 75% of that of the mAb P2-284-treated group at 145 dpi (Fig. 4c), immunohistochemical analysis revealed little difference in PrPSc accumulation in the hippocampi, thalami or cerebellums of these mAbs-treated groups (Fig. 5, PrPSc). Quantitative analysis of microglial activation, astrogliosis, and vacuolation showed no significant differences between mAb 31C6- and negative control mAb-treated mice in the hippocampi or thalami (Fig. 5). However, there were significantly fewer Iba-1-positive microglia in the cerebellums of mAb 31C6-treated mice than in those of the negative control mAb-treated mice. Consistent with the decreased microglial activation, astrogliosis in the cerebellums of the mAb 31C6-treated mice appeared to be slightly milder than in the mAb P2-284-treated mice. In addition, vacuolation in the cerebellums of the mAb 31C6-treated mice was also reduced to 40% of that of the mAb P2-284-treated mice; however, these differences were not statistically significant due to the presence of large variations.
Because antibodies do not usually diffuse from blood vessels into the brain across the BBB, anti-PrP antibody administered via a peripheral route would be expected to distribute inefficiently to the brain parenchyma . However, the brain distribution of anti-PrP antibodies administered via peripheral routes, including the i.v. route, has not been well characterized. In this study, we showed that an anti-PrP antibody administered via the tail veins of Jcl:ICR mice during the clinical phase of infection with the Chandler strain were detectable in their cerebellums and thalami as the disease progressed (Fig. 1). Vorbrodt et al. have reported that the permeability of BBB in cerebellum, but not the hippocampus or cerebral cortex, increases in MB mice during the clinical phase of infection with the prion 87V strain . Actually, several studies have revealed functional impairment of the BBB in prion-infected animal models [21-23]. In contrast, researchers did not observe extravasation of immunoglobulin in the cerebellums of tga20 mice infected with the Chandler strain at any stage of the disease, including the clinical phase, suggesting that BBB permeability does not increase with progression of this disease . Interestingly, during experimental infection with rabies virus, the permeability of the BBB was reportedly more strongly increased in relation to the cerebellum than to the cerebral cortex, even though the viral loads in each region were comparable . Therefore, the degree of impairment of the BBB likely depends on the particular prion and mouse strain used as well as the region of the brain involved. Thus, our observation that mAb 31C6 distribution to the cerebellar medulla and thalami appeared to be more efficient when we administered the mAb on 140 dpi than when we administered it on 120 dpi suggests that, in our model, BBB permeability does increase with disease progression.
To date, there have been no reports that peripheral administration of anti-PrP antibodies during the clinical stage prolongs survival of prion-infected mice. Here, we showed that more than half of Chandler strain-infected mice that had received serial injections of mAb31C6 via their tail veins from 120 dpi survived longer than the Chandler strain-infected mice that had received the negative control mAb (Fig. 2). The mAb 31C6 appeared to be less efficiently distributed in the cerebellums and thalami of the mice whose survival times were not prolonged than in those of the mice that survived longer than the mean survival time of the negative control group (Fig. 3). In mice in the late stage of infection with the Chandler prion strain, severe emaciation makes it difficult to inject mAb into the tail veins. Thus, we cannot exclude the possibility that the inefficient delivery of mAb that we identified was due to technical difficulties with i.v. injections. Interestingly, in later experiments, in which i.v. injections were performed by the same person as in the experiment shown in Figure 2a, the mice that received mAb 31C6 had significantly smaller decreases in body weight than did the mice that had received negative control mAb at 141 dpi (Fig. 4a). This was true even when we compared all six mice in each group. This could be attributable to the researcher more successfully accessing the tail veins of emaciated, late stage mice experience after gaining considerable experience in administering repeated i.v. injections in our earlier experiments. Therefore, combining data shown in Figure 2 with that in Figure 4a, our findings suggest that administration of an anti-PrP mAb through a peripheral route could slow disease progression in mice infected with the Chandler strain.
Our previous study showed that brain regions that displayed milder vacuolation, astrogliosis and microglial activation than that seen in mice infused with negative control mAb were the same regions as those in which anti-PrP mAb was well distributed following intraventricular administration . Moreover, delivery of genes encoding scFv or PrP-Fc fusion protein reduce the PrPSc burden around the delivery site [14, 15]. In the present study, i.v. administration of mAb 31C6 reduced vacuolation and microglial activation in the cerebellum; however, we observed little effect in the hippocampus or thalamus (Fig. 5). These findings suggest that the amount of antibody distributed in the cerebellum after i.v. administration is sufficient to mitigate the progression of neurodegeneration caused by prion propagation.
The relationship between functional impairments and affected brain areas is poorly understood . The intriguing question of whether attenuation of neuropathological lesions directly prolongs survival remains unanswered. Resolution of this question may lead to identification of clinical target sites of prion infection. Researchers have observed diffused PrPSc deposition in the cerebral cortexes, hippocampi, thalami, cortexes and cerebellar medulla of mice infected with the Chandler strain [13, 27]. Deceleration in PrPSc accumulation and slowing of progression of neurodegeneration in the hippocampi and thalami after intraventricular infusion of anti-PrP mAbs implies that specific regions in the hippocampi or thalami are some of the clinical target sites in Chandler strain-infected mice . Here, we showed that i.v. administration of anti-PrP mAb prolongs the survival of mice infected with the Chandler strain and slows progression of neurodegeneration, as evidenced by vacuolation, astrogliosis or microglial activation, in their cerebellums, but has little effect in their hippocampi or thalami (Fig. 5). These data suggest that the hippocampi and/or thalami may not be unique clinical target sites in mice infected with the Chandler strain. Indeed, because we examined only two coronal levels in both studies [approximately levels 48 and 84 in “The Mouse Brain Atlas” ], extensive analysis of other regions is required to identify clinical target sites more accurately.
Because mAbs 31C6 and P2-284 share the same isotype, IgG1, there is little difference in their size or structure. Therefore, if mAbs spread by passive diffusion because of increased permeability of the BBB, comparable amounts of control and anti-PrP mAbs should enter the brain parenchyma. However, we detected little mAb in mice cerebellums or thalami when we serially administered mAb P2-284 via the tail veins. In contrast, in mice that had received mAb 31C6, mAbs were clearly present in the brain parenchyma and the total level of PrP was increased (Figs. 1 and 4). We observed a similar difference in antibody distribution following intraventricular infusion of anti-PrP mAb; mAb 31C6 infused into the lateral ventricle was still detectable in the hippocampus more than 2 weeks after termination of the infusion, whereas we detected only a low level of mAb P2-284 in the same region, even during infusion. In addition, we observed increased total PrP in mice intraventricularly infused with mAb 31C6 . The increased level of PrP was probably due to retention of antibody-PrPC complexes on the cell surfaces [3, 5]. Therefore, in mice that had received i.v. mAb 31C6, in addition to the presence of mAb 31C6 in the brain parenchyma (Figs. 1 and 2), total PrP was increased but of PrPSc decreased (Fig. 4), suggesting that mAb 31C6 remains in the brain parenchyma as an antibody-PrPC complex.
Binding of anti-PrP antibodies to PrPC can inhibit the interaction between PrPC and PrPSc directly or indirectly . Antibody may directly block the domain on PrPC that is required for interacting with PrPSc; inhibition of PrPSc formation by Fab or scFv fragment suggests that this inhibition is a direct effect [6, 15, 29-31]. Alternatively, retention of the antibody-PrPC complex may reduce the rate of entry of PrPC into the endocytic pathway, which is reportedly required for PrPSc formation [32, 33]. Thus an inadequate supply of PrPC as a substrate for PrPSc formation in cellular compartments, where the conversion of PrPC into PrPSc takes place, may indirectly inhibit PrPSc formation. I.v. administration of mAb 31C6 slows PrPSc formation but increases the total PrP, implying that alteration of the normal PrPC metabolic pathway by capturing PrPC with mAb 31C6 is one of the mechanisms by which mAb 31C6 inhibits PrPSc formation.
This study shows that passive immunization through a peripheral route may be useful for the treatment of prion diseases. Progression of neuropathological lesions was mitigated in the cerebellum, where anti-PrP mAb was well distributed. However, the mAb was not well distributed to the hippocampus or thalamus, suggesting that improvement of mAb delivery will greatly enhance the therapeutic effect of anti-PrP antibody.
This work was supported by a Grant-in-Aid for Science Research (A) (Grant no. 23248050) and for challenging Exploratory Research (Grant no. 23658233), a grant from the global COE Program (F-001), the Program for Leading Graduate Schools (F01), and the Japan Initiative for Global Research Network on Infectious Diseases (J-GRID), from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. This work was also supported by grants for TSE research (H23-Shokuhin-Ippan-005) and Research on Measures for Intractable Diseases from the Ministry of Health, Labor and Welfare of Japan. This work was also partly supported by a Grant-in-Aid from the BSE Control Project of the Ministry of Agriculture, Forestry and Fisheries of Japan. We also thank Zensho for the BSL3 facility.
The authors declare no conflict of interest or financial support.