. Amyloid imaging in distinguishing atypical prion disease from Alzheimer disease. Neurology. 2007 Jul 17;69(3):283-90. PubMed.

Recommends

Please login to recommend the paper.

Comments

  1. The paper from Boxer et al. compares two brothers (35 and 41 years old) who were ultimately found to be concordant for a 6-octapeptide repeat insertion mutation (6-OPRI) in the gene that codes for the prion protein (PrP). The younger brother was initially diagnosed with early onset Alzheimer disease (AD) after a 3-year history of progressive memory loss, and both brothers were said to have “similar clinical syndromes.” Both had similar FDG scans judged to be typical for AD because of the regional pattern of cerebral hypometabolism. Their father had died at age 47, apparently without a confirmed diagnosis, after a 10-year history of progressive dementia. The authors point out that since there is considerable overlap in the clinical presentations of early onset dementia (EOD) caused by prion and AD gene mutations, a major motivation for this study was to determine if amyloid imaging could be helpful in the differential diagnosis of EOD. The younger brother underwent amyloid imaging with FDDNP (Shoghi-Jadid et al., 2002) at UCLA, and the older brother was imaged with PiB (Klunk et al., 2004) at the Lawrence Berkeley National Laboratory.

    The 35-year-old brother’s FDDNP scan was read as “intermediate” between controls and AD, with asymmetric (right > left) FDDNP retention. The highest FDDNP values were in the medial temporal and parietal lobes. The FDDNP retention pattern was noted to be of “lesser intensity and in a pattern different from that typically found in AD.” The 41-year-old brother had a control-like (i.e., normal) pattern on his PiB scan.

    The conclusion drawn from the authors’ review of the two imaging studies was that “Although not performed in the same subject, the PIB and FDDNP results suggest that FDDNP has a greater binding affinity for PrP-amyloid than PIB.” However, when one critically considers the data presented, along with the known neuropathology of this PrP gene mutation, one must question whether the FDDNP signal relates to PrP-amyloid deposition at all. One also must ask if the addition of either amyloid imaging scan to the clinical workup would have aided a clinician in making the differential between early onset AD and a prion disease.

    First, consider the data from this review paper. The data in Figure 3 of Boxer et al. were presented on different scales, although the same units were used for each tracer (i.e., global distribution volume ratio—DVR). If re-plotted on the same scale (Figure A), one sees that the dynamic range of the FDDNP data is quite small (~15 percent of PiB in this study). When the PiB data are expanded in the same way as the FDDNP data, one also can see that the relationship between the 6-OPRI patients and the controls is nearly the same for both tracers (except for the one higher control in the PiB group). This is not to say that PiB is detecting prion deposits in the 41-year-old brother, but that perhaps some non-specific abnormality in the 6-OPRI patients skews them toward the higher end of controls with both agents. In other words, it may be that neither tracer is specifically detecting PrP-amyloid in this study.

     

    image

    Figure A. FDDNP and PiB data modified from Figure 3 of Boxer et al. Both datasets are now plotted on the same graph scaled as Boxer et al. had scaled the PiB data (gray graph) or the FDDNP data (white inset).

    A more significant indication that neither tracer is specifically binding to PrP-amyloid is the discordance between the regional pattern of FDDNP (and PiB) retention described in this study and the known regional distribution of PrP-amyloid pathology. This is true both for 6-OPRI kindreds in general, and, more importantly, for the postmortem pathology of the 41-year-old brother, who died while the Boxer et al. paper was being written. The authors point out that previous studies showed 6-OPRI kindreds to have variable cortical spongiosis with or without diffuse PrP-amyloid deposition in the cerebellar molecular layer (Vital et al., 1998). More pertinent information comes in the “Note Added in Proof.” Here, the authors state that “Postmortem PrP immunohistochemical staining demonstrated the presence of prominent PrP-amyloid deposition in the cerebellar greater than in the cerebral cortex in Patient 1” (i.e., the 41-year-old brother). The pathology clearly indicates that if FDDNP were binding to PrP-amyloid pathology, the signal should be highest in the cerebellum. Not only was increased FDDNP retention not observed in the cerebellum, but the retention in the cerebellum was so low that the authors used this region as their reference tissue. Assuming that the postmortem findings in the 41-year-old brother are reflective of the in vivo pathology in the brain of the 35-year-old brother, one must question the biological basis of FDDNP binding in this patient. This also leaves us with considerable uncertainty as to how we should interpret the in vivo FDDNP data. The authors recognize the mismatch between the distribution of PrP-amyloid pathology and regional FDDNP retention with the statement, “Weak binding of FDDNP to cortical spongiform pathology might also potentially play a role in this subject because spongiosis is common in 6-ORPI.” This statement seems to say that FDDNP retention is not determined by direct and specific binding to PrP-amyloid, but may reflect the spongiform changes as well. This is certainly not the conclusion that is emphasized in this paper.

    In the end, we must come back to the ultimate motivation for imaging these subjects: clinical diagnosis. The biological basis of FDDNP or PiB retention may not matter if amyloid imaging with either tracer can help to accurately distinguish between two serious diseases with vastly different pathologies and genetic causes. If effective therapies become available for either of these dementias in the future, they are likely to be very different as well, so diagnosis will become even more important. Would the amyloid imaging studies have aided in the accurate diagnosis of these patients? In the setting of this study, the patients presented with obvious clinical dementia. The goal of imaging would be to determine the underlying etiology of the dementia, not to decide whether the patients were normal or abnormal. One can use the data from Table 2 of Boxer et al. to prepare a mock “clinical report” (Figure B).

     

    image

    Figure B. Mean ± SD for FDDNP (left) and PiB (right) DVR vales using cerebellum as reference produced from Boxer et al. Table 2. For clarity, only the right-sided data is shown. Abbreviations: “R-” indicates right-sided data; Fr: frontal; MT: medial temporal; LT: lateral temporal; Par: parietal; PC: posterior cingulate. (Triangles: controls; circles: AD patients; squares: patients.)

    A clinician receiving the FDDNP report would see that the 35-year-old brother’s FDDNP retention is AD-like in most brain regions—although the result falls in or near the upper range of control values in many regions as well. The clinician would likely decide that this is an abnormal scan. However, since FDDNP is reported to bind to both plaques and tangles, and now to PrP-amyloid pathology and spongiform changes, the clinician would again be left wondering what caused the dementia. In other words, the clinician would gain little information from the FDDNP scan that the clinical presentation hadn’t already provided. The clinician receiving the PiB report would see that the 41-year-old’s PiB retention fell within a narrow normal range in all brain areas except medial temporal—an area in which AD and control subjects show a large overlap. This clinician would likely conclude that the etiology of the patient’s dementia is not based on Aβ amyloid deposition (i.e., not likely to be AD). Because PiB appears to detect primarily, if not exclusively, Aβ deposition in vivo, a negative PiB result can only rule out Aβ deposition and therefore rule out AD (Klunk et al., 2003; Klunk et al., 2005). It cannot point to any other pathology or etiology. However, the clinician has now excluded a major possibility from the differential diagnosis and can focus on the other etiologies for EOD. Excluding AD from the differential diagnosis would most likely lead to a search for other causes of EOD such as the abnormalities in the 6-OPRI mutation that eventually proved to be the cause of this familial dementia.

    References:

    . The binding of 2-(4'-methylaminophenyl)benzothiazole to postmortem brain homogenates is dominated by the amyloid component. J Neurosci. 2003 Mar 15;23(6):2086-92. PubMed.

    . Binding of the positron emission tomography tracer Pittsburgh compound-B reflects the amount of amyloid-beta in Alzheimer's disease brain but not in transgenic mouse brain. J Neurosci. 2005 Nov 16;25(46):10598-606. PubMed.

    . Imaging brain amyloid in Alzheimer's disease with Pittsburgh Compound-B. Ann Neurol. 2004 Mar;55(3):306-19. PubMed.

    . Localization of neurofibrillary tangles and beta-amyloid plaques in the brains of living patients with Alzheimer disease. Am J Geriatr Psychiatry. 2002 Jan-Feb;10(1):24-35. PubMed.

    . Prion encephalopathy with insertion of octapeptide repeats: the number of repeats determines the type of cerebellar deposits. Neuropathol Appl Neurobiol. 1998 Apr;24(2):125-30. PubMed.

Make a Comment

To make a comment you must login or register.