bioRχiv hosts two much-awaited manuscripts characterizing new tau PET tracers in people with Alzheimer’s disease. One study, led by Tobey Betthauser of the University of Wisconsin-Madison School of Medicine, tested Merck’s candidate, MK-6240. The other, co-led by Dean Wong of Johns Hopkins University in Baltimore and Edilio Borroni at Hoffman-La Roche in Basel, Switzerland, compared three Roche contenders, ultimately zeroing in on one—RO-948—as the best choice. This paper was published May 4 by the Journal of Nuclear Medicine.

  • Tau tracers MK6240 and RO-948 are being tested in cognitively normal controls and people with AD.
  • In the latter, both ligands efficiently labeled regions associated with neurofibrillary tangle pathology.
  • Both had minimal non-specific binding in the choroid plexus and basal ganglia, but did bind the meninges and some other off-target regions.

Both manuscripts report that the tracers efficiently entered the brain and, in people with AD, labeled regions known to accumulate neurofibrillary tangles. They report no off-target binding of their respective tracers in the choroid plexus, a region where nonspecific uptake of Avid’s AV1451/flortaucipir has been said to potentially obscure signals from nascent tau pathology in the nearby hippocampus. However, both companies' tracers still bound where tau aggregates feared to tread, namely in the meninges. Whether either of the new tracers will supplant flortaucipir, or simply enlarge the pool of available options, is uncertain.

“The importance of tau imaging coupled with the shortcomings of first-generation tau agents, particularly with respect to off-target binding and subcortical uptake, has led to the generation of new imaging agents,” commented Andrew Stephens of Piramal in Berlin. “Undoubtedly these new noninvasive imaging biomarkers will help us to understand the underlying mechanism of dementia and identify disease-modifying treatments.” Piramal is developing its own tau tracer, and Stephens wrote that larger clinical studies will be needed to compare tracers.

Postmortem pathology studies have shown that neurofibrillary tangles spread throughout the brain in a characteristic pattern in people with AD, starting in the medial temporal lobe and eventually fanning out across the neocortex (Braak and Braak, 1991). The advent of tau PET tracers has allowed researchers to track this in living people. Flortaucipir was the first to be widely used. While it robustly labels regions of the brain riddled with tau pathology in people with AD, it also binds other areas, notably the choroid plexus and substantia nigra (Feb 2015 newsFeb 2016 conference news), where neuropathology says there are no tangles in AD. Binding in the choroid plexus can spill over into the nearby medial temporal lobe, complicating efforts to pick up the first inklings of tau pathology in affected individuals. The tracer also doesn’t effectively label tau isoforms in tauopathies other than AD (Feb 2016 conference news).

Will second-generation tracers fare better? A handful are wending their way through development, Merck’s MK-6240 being a frontrunner (Apr 2017 conference newsDec 2017 conference news). It has high affinity for neurofibrillary tangles, and preclinical studies in nonhuman primates as well as a small Phase 1 trial in people, revealed no off-target binding in worrisome regions including the choroid plexus (Hostetler et al., 2016Apr 2016 newsApr 2017 news).

For the current study, posted on bioRχiv March 28, Betthauser and colleagues tested MK-6240 in 51 people recruited from the Wisconsin-Madison AD Research Center or the Wisconsin Registry for Alzheimer’s Prevention. The cohort included three young controls from 27 to 45 years old, 33 controls between 56 and 77, two people with mild cognitive impairment, and seven with a clinical diagnosis of probable AD. The group also included six “cognitive decliners,” people who were initially cognitively normal, but whose memory slipped between multiple clinical visits. All participants, except the young controls, also had PiB-PET scans to assess Aβ accumulation. Twenty-two people carried the ApoE4 allele.

The researchers acquired MK-6240 from Cerveau Technologies, a Boston-based company licensed by Merck to manufacture and sell the tracer. Cerveau has been forging agreements with numerous PET groups across the AD field to evaluate MK-6240, most recently with MGH and the University of Pittsburgh (Dec 2017 conference news and company press releases).

To narrow down an optimal time window to image brain uptake, the researchers dynamically scanned 19 of the participants for an extended duration of 1.5 to two hours after injecting MK-6240, tracking its rise and fall throughout the brain and other tissues of the body. Using the cerebellum as a reference region, they found that the standardized uptake value ratio (SUVR) peaked between 70 and 90 minutes for most regions known to harbor tau pathology. In contrast, in regions with off-target binding, including bone marrow and meninges, uptake continued to rise throughout the two-hour scan. Based on this and further analyses, the researchers settled on the 70-90 minute time window as the best to catch the strongest, most specific signal for this tracer.

During this timeframe, the researchers compared SUVRs across multiple regions. Except for one person who had an unusual distribution of tracer uptake, everyone who tested negative for Aβ via PiB-PET also took up no MK6240 in Braak regions, according to Betthauser. Among Aβ-positive people, the researchers observed tau distributions that spanned the entirety of Braak staging: Some people had tracer uptake confined to the medial temporal lobe (Braak stage I), while in others, uptake extended into the inferior temporal cortex and beyond.

Spanning the Tau Spectrum. Among people who tested positive for PiB-PET, MK-6240 uptake ranged from extensive neocortical uptake (left) to entorhinal uptake only (right) (Dem=dementia, OC=older control, CD=cognitive decliner). [Courtesy of Betthauser et al., bioRχiv, 2018.]

Ligand binding to non-tau areas varied in magnitude across participants. The ligand bound the bone marrow, ethmoid sinus, pineal gland, substantia nigra, superior anterior vermis, superior cerebellum, and the meninges. In six extreme cases, uptake in the meninges appeared to spill over into nearby cortical regions. No binding was observed in the basal ganglia; one person had weak binding in their choroid plexus.

The authors concluded that MK-6240 kinetics were similar to flortaucipir’s. Unlike flortaucipir, MK-6240 had no off-target binding in regions near the entorhinal cortex or hippocampus, thus theoretically allowing for detection of emerging tau pathology in those important bellwether regions.

Betthauser and co-author Sterling Johnson told Alzforum that the findings so far support the idea that Aβ accumulation precedes tau deposition, as all participants with MK-6240 uptake in Braak regions were also Aβ-positive, but not vice versa. The scientists are expanding their studies to include more participants and to run longitudinal scans, they said.

Roche Enters the Ring
The Roche paper posted on bioRχiv on April 11. Co-first authors Wong and Robert Comley, previously at Hoffman-La Roche and now at AbbVie, compared the pharmacokinetic properties of RO-963, RO-643, and RO-948. Previously, these chemically similar compounds were shown to bind neurofibrillary tangles in postmortem brain slices, and to have suitable pharmacokinetic properties in mice and nonhuman primates (Feb 2015 newsGobbi et al., 2017Honer et al., 2017).

For the current study, Wong and colleagues injected the tracers into seven healthy controls aged 25 to 40, and seven older people with AD. To avoid the confounds of potentially uneven distribution of brain tau among such a small cohort, each participant received two tracers, with a washout of one to two weeks between scans. In the young controls, all three ligands entered the brain quickly, but RO-948 was taken up the most. While RO-643 and RO-948 quickly washed out of the brain in the controls, RO-963 lingered, suggesting non-specific binding. In people with AD, RO-948 signals in predicted tau-rich regions better contrasted those in tau-poor regions than did RO-643 signals. RO-948 also better distinguished between people with AD and young controls, hence RO-948 made the cut.

The researchers collected arterial blood from a subset of 11 controls and 11 AD patients for a distrubution volume ratio (DVR) calculation, i.e., the tracer amount in any region of interest compared to that in the blood at a given time after injection. This measurement is the gold standard for PET studies, but it requires long scan times, and an invasive procedure to sample blood repeatedly over time. The researchers found that SUVR measurements taken between 60 and 90 minutes after injection aligned closely with DVR measures and selected that as the sweet spot to measure SUVR. RO-948 aligned most closely with the DVR measurements during this timeframe.

They next compared RO-948 uptake among 22 regions of the brain, including regions implicated in Braak staging. SUVRs from five controls who averaged 62 years old and 11 people with AD who averaged 64 years old indicated that the AD patients had more tau in 13 regions, including five in which the lowest values in people with AD exceeded the highest values from controls: the right entorhinal cortex, left and right parahippocampi, left fusiform gyrus, and left middle temporal cortex.

RO-948 Hotspots. Surface projection maps of mean RO-948 SUVR uptake among 11 AD subjects. [Courtesy of Wong et al., bioRχiv, 2018.]

Using the uptake patterns, the researchers classified three of the people with AD as Braak stage IV, two as stage V, and six as stage VI, but this in vivo staging did not correlate with MMSE in this small cohort. In four of these people, the researchers did a follow-up scan between six months and 21 months later, finding that tau deposition had markedly increased in some regions in three of them.

The researchers also saw RO-948 bound the substantia nigra, the cerebellar vermis, and the meninges but not in the choroid plexus or striatum. While the lack of off-target labeling in these latter two regions is welcome, Wong told Alzforum that future studies in larger cohorts will be needed to fully rule out nonspecific binding there.

Overall, the researchers contend that its high specificity for tau pathology, absent off-target labeling in the choroid plexus, and potential for reliable and consistent production of the tracer make RO-948 a good candidate for longitudinal, multicenter studies; however at this time it is only being used academically. Similar to claims made by Betthauser and colleagues about MK-6240, Wong and colleagues proposed that RO-948 might pick up early changes in tau pathology in the medial temporal lobe, although further longitudinal studies, particularly in people in earlier stages of the disease, will need to be conducted to find out.

Keith Johnson of Massachusetts General Hospital in Boston called both studies a welcome addition to the field, but was not entirely convinced by the benign nature of the off-target labeling. The tracers’ binding in the meninges, which increases throughout the scan, could interfere with accurate detection of early stage tau pathology in nearby brain regions, he noted. This would be problematic in the context of longitudinal studies, if serial PET scans don’t all measure the exact same span of the kinetic curve, he wrote (see comment below).—Jessica Shugart


  1. These papers are welcome additions to the literature. The Wong et al. paper exemplifies more recently developed tau radiopharmaceuticals, and the Betthauser et al. paper reaffirms the need for thorough evaluations of newer agents by independent academic centers.  A couple things struck me. First, late-stage, catastrophic tau deposition is now readily detectable with PET as very high signal in the expected neocortices. This is a tremendous success for the field, but for AD prevention trials we need to detect reversal of early-stage tauopathy where weaker signal emerges from medial temporal regions. Kinetic properties need to be carefully demonstrated with focus on serial measures in early stages.

    Second, leptomeningeal off-target binding is evident in tau PET tracers, including in these two reports. We will need autoradiography data as well as postmortem correlations to clarify the exact substrate. These two problems are related: meningeal signal that rises in later frames (see Betthauser et al., Figure 2) could spill into brain and obscure early tau change when serial PET acquisitions fail to measure the same span of the kinetic curve at followup. The new papers move us closer to overcoming these barriers and help clarify the path forward.

    Finally, we deserve better from the Journals. The technology used to report our image data has not changed in decades. Only a tiny fraction of the image data supporting the science is available for reviewers and readers to judge. We need to see entire image data volumes, not the 2 percent sample provided, so we can evaluate entorhinal, substantia nigra, retina, choroid plexus, signal dynamics, etc., etc.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Tau Tracer T807/AV1451 Tracks Neurodegenerative Progression
  2. Shaky Specificity of Tau PET Ligands Stokes Debate at HAI
  3. At HAI, Researchers Explore Diagnostic Potential of a Tau Tracer
  4. Next-Generation Tau PET Tracers Strut Their Stuff
  5. At CTAD, Tau PET Emerges as Favored Outcome Biomarker for Trials
  6. Improving Tau PET: In Search of Sharper Signals
  7. Human Amyloid Imaging Meeting Was Abuzz With Talk of Tau

Paper Citations

  1. . Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 1991;82(4):239-59. PubMed.
  2. . Preclinical Characterization of 18F-MK-6240, a Promising PET Tracer for In Vivo Quantification of Human Neurofibrillary Tangles. J Nucl Med. 2016 Oct;57(10):1599-1606. Epub 2016 May 26 PubMed.
  3. . Identification of Three Novel Radiotracers for Imaging Aggregated Tau in Alzheimer's Disease with Positron Emission Tomography. J Med Chem. 2017 Sep 14;60(17):7350-7370. Epub 2017 Jul 12 PubMed.
  4. . Preclinical Evaluation of18 F-RO6958948,11 C-RO6931643 and11 C-RO6924963 as Novel Radiotracers for Imaging Aggregated Tau in AD with Positron Emission Tomography. J Nucl Med. 2017 Sep 28; PubMed.

External Citations

  1. Phase 1 trial
  2. company press releases

Further Reading

No Available Further Reading

Primary Papers

  1. . In vivo characterization and quantification of neurofibrillary tau PET radioligand [18F]MK-6240 in humans from Alzheimer's disease dementia to young controls. bioRχiv 290064; Apri 25, 2018
  2. . First in-human PET study of 3 novel tau radiopharmaceuticals: [11C]RO6924963, [11C]RO6931643, and [18F]RO6958948. J Nucl Med jnumed.118.209916 published ahead of print May 4, 2018