Glutamine-based PET imaging facilitates enhanced metabolic evaluation of gliomas in vivo

 

 

Craig B. Thompson et. Al

Link to the paper

Screen Shot 2015-03-27 at 11.05.38 PM

Generally, cancer cells contain some metabolic mutation that favors rapid reproduction and the accumulation of biomass, and subsequently results in an increased uptake of both glutamine and glucose. In this paper, the authors further subcategorized cancer cells into primary glibolastomas and secondary glioblastomas. Primary glioblastomas tend to show deregulation of PI3K, which results in unbalanced growth factor throughout the cell. Secondary glioblastomas, on the other hand, tend to have mutations in Isocitrate dehydrogenase 1 (IDH1), which results in the accumulation of the oncometabolite 2-hydroxyglutarate (2-HG). The rapid uptake of glucose has traditionally been used to diagnose and quantify different cancers through the use of 18F-FDG PET imaging (where 18F-FDG is 18F-fluorodeoxygluocse, a fluorescing analog of glucose, and PET is positron emission tomography). This technique, however, is less effective at diagnosing cancer in the brain due to the high glucose metabolism naturally present in the brain and results in poor imaging and therefore poor diagnosis. Current imaging analogs other than 18F-FDG, such as amino acid analogs, while more effective at diagnosing brain tumors, are still incapable of indiscriminately determining the metabolic state of glioblastomas.  Given the high fatality rate of “neurologically destructive gliomas”, it is clear more noninvasive forms of diagnosis that allow for greater characterization of the metabolic state of the tumor need to be developed.

The authors of this paper, hypothesized, that the rapid uptake of glutamine in cancer cells (cancers such as lymphoma and pancreatic adenocarcinoma use increased glutamine metabolism to support growth), but not in normal brain cells, might offer an avenue to achieve this new method of diagnosis. Previous to this paper, the authors developed a fluorescent analog to glutamine, which they labeled 18F-FGln (4-18F-(2S, 4R)-fluoroglutamine). In this paper, the authors showed that brain tumors demonstrated high uptake of the analog, whereas normal brain cells showed minimal uptake, which suggests that this molecule in combination with PET imaging can be used to effectively diagnose and distinguish between different brain cancers. The authors also demonstrated that 18F-FGln allowed for greater delineation and assessment of the metabolic uptake of gliomas, suggesting it may be used to more effectively determine the progression of glioblastomas.

 

So how did the authors go about proving their hypothesis? First, the authors proved that 18F-FGln was a viable substrate to use for PET imaging. To accomplish this, they compared the metabolism of glutamine with the metabolism of 19F-Gln (a more stable form of 18F-FGln, which allowed for mass spec analysis), and discovered that cells given normal glutamine produced glutamate, however, cells given 19F-Gln produced no 19F-glutamate. This signified to the authors that the fluorescence witnessed during PET imaging corresponds strongly to glutamine uptake and that 18F-FGln is trapped in gliomas after uptake by incorporation into protein. The authors then injected mice with with various types of glioma cells to create xenograft (the cells that are transferred from one organism to another) models. The authors then injected both the control mice and the xenograft models with 18F-FGln and 18F-FDG and used PET scans to generate images of the various brains. The figure below discusses some of their results. Part A of the figure simply shows the structure of 18F-FGln. Figure B shows the amount of 18F-FGln taken up by a normal brain as opposed to the xenograft models. Figure C shows the amount of 18F-FDG taken up by the normal brain as compared to the xenograft models. The take home message here is that the normal brain took up significantly less 18F-FGln compared to the xenograft models than 18F-FDG. Figures E, F, and G are PET scans of mice with normal brains, and two different xenograft models, respectively. Figure E demonstrates that there is clearly more background uptake of 18F-FDG than 18F-FGln. Figures F and G amplify this point, as there is clearly much higher tumor resolution in the mice injected with 18F-FGln as opposed to the mice injected with 18F-FDG, as indicated by the red arrows and the defined level of substrate uptake (red being the highest) in 18F-FGln injected mice.

Screen Shot 2015-03-27 at 10.55.03 PM

 

 

In the next section of the paper, the authors attempted to address any factors that might affect the quality of the image produced. Because 18F-FDG PET scans have historically given poor scans during inflammation, the authors created “robust neuroinflammation” within mice models by injecting the mice with immune stimulating macrophages and interferon-γ (both of which result in an inflammatory phenotype). The authors then produced PET scans using both 18F-FGln and 18F-FDG, and determined that while there was increased background uptake of 18F-FDG during inflammation, there was no increased background uptake of 18F-FGln during inflammation. As increased background uptake results in poorer resolution, this data suggests that 18F-FGln might surpass 18F-FDG as a diagnostic tool in the presence of inflammation, especially in the brain.

 

In their last sections of the paper, the authors correlated 18F-FGln uptake with the metabolic stage and robustness of tumors. The authors demonstrated that a mouse model treated showed significantly less uptake of 18F-FGln after treated with chemotherapy (this was confirmed by PET scans and MRI imaging). The authors’ subsequent study might be their most significant, as they shifted their focus to human trials. First, the authors determined using siRNA knockdown and pharmacological inhibition that increased expression of SLC1A5 (a glutamine transporter) in gliomas as compared to normal brain tissue correlated to increased uptake of 18F-FGln, suggesting to the authors that the substrate might have wider applications in characterizing gliomas than just the genotypes tested. Finally, the authors demonstrated that 18F-FGln results in high uptake in progressive tumors in humans, and can potentially be used to help identify transforming tumors.

 

This paper is interesting in terms of significance as it seems the authors are attempting to occupy a niche among all the other imaging analogs that have been released and used in the past few years. The significance, from the authors standpoint, is that with more development and research, hopefully 18F-FGln can confidently discriminate between various stages of cancer, and provide the doctor with an accurate metabolic characterization of the glioblastoma. Obviously correct and encompassing characterization of cancer is essential for timely diagnosis and treatment, so in that respect, if the authors eventually succeed in proving that this analog can without a doubt determine the metabolic state of gliomas, then this analog might fortify its niche within the field of cancer diagnosis.

 

By the way, I could not figure out how to hyperlink in the comments, so here are the links to the papers I cited in the comments:

1. Synthesis of Optically Pure 4-Fluoro-Glutamines as Potential Metabolic Imaging Agents for Tumors

2. Glutamate-induced neuronal death: A succession of necrosis or apoptosis depending on mitochondrial function

3. Molecular imaging of gliomas with PET: Opportunities and limitations

14 Replies to “Glutamine-based PET imaging facilitates enhanced metabolic evaluation of gliomas in vivo”

  1. Hey Tommy! I hate to be the bearer of bad news, but the link you posted doesn’t give us access to the article. On the other hand, I enjoyed your summary and am confident that you outlined the most important details for us.

    I agree with you that the direct applicability of this new-found method and their experimentation in vivo significantly escalated the impact of this paper.

    My questions have to do with the radioactive glutamine uptake. This method is effective because it deviates from the traditional glucose PET scans and the cancer cells uptake glutamine without metabolizing it. Are you saying that there is no metabolism of glutamine by cancer cells or that it is simply not quick enough to have excreted glutamine by the time the PET scan is taken? I’m wondering if there could be a concentration in which glutamine could build up in cells and act as a toxin to the brain?

    My second question is a little simpler.. You said that radioactive glutamine uptake is dependent on ‘the metabolic stage and robustness of tumors.’ Are gliomas always apt to uptake glutamine at an observable level, or are there detection limits early on in the tumor formation where they still can’t be recognized?

    Good work on the summary!

    1. Hey Ryan

      To address your first question, I don’t think the authors were saying that there is no metabolism of glutamine by cancer cells, but just that the glioblastomas in question don’t seem to metabolize their fluorinated glutamine analog.

      I never thought about your first question when I was reading the paper, but now that you bring it up, it is an interesting thought. I tried looking up glutamine as an exitotoxin and I could not seem to find anything. What I did find, however, is that glutamate is a potent endogenous and exogenous exitotxin and excess concentrations can lead to a cellular cascade that eventually results in apoptosis of neurons (Pierluigi et. Al). So I don’t think the excess glutamine has a danger in resulting in neuron death, but now that I think about it, I could imagine that even though the authors did not explicitly state this in their paper, glutamate poisoning of patients might have been an underlying reason why the authors decided to test if their analog was metabolized to glutamate.

  2. Hey, Tommy. I also could not get access to the paper through the link that you provided, but I will trust your summary. I feel as though this is a good example of a translational medicine paper and uses the concepts we have learned in class in a clinical setting. From what we have discussed in this unit, I did not find these results too surprising. The fact that cancer cells dysregulate primary metabolism to supplement the accumulation of biomass through anaplerosis is a concept that we have seen throughout this unit, and this paper provides another real-world example of this taking place. I know that you mentioned that primary gliomas perform increased amino acid catabolism to supplement their carbon needs, but in your analysis, it wasn’t clear whether the authors actually tested primary gliomas. Can you clarify this? I would look up the paper if I had access to it.

    Also, I wonder whether a fluorescent tag could be developed to attach to 2HG. What do you think? It might be difficult because 2HG is so close to alpha KG structurally, but that could be interesting. It might also be useful for more genetic screenings for mutations in the IDH gene to take place to determine if patients could be more susceptible to various types of cancer. Just some thoughts. Thanks!

    1. Hey Zach,

      To answer your first question, the authors did in fact derive two models to test both primary and secondary gliomas. For primary gliomas, they used platelet-derived growth factor model “to medol primary GBMs with enhanced PI3K/AKT/mTOR signaling”. To model secondary gliomas, the authors used IDH1 mutants.

      For your second question, I think designing a fluorescent tag for 2-HG might be an interesting and noninvasive way to determine the nature of a glioblastoma. Although the structure of 2-HG and alpha-keto gutarate are very similar, I think if somehow you could design a tag that could selectively bind to just either of these two molecules, determining whether or not a mass resembled a cancer cell would probably be rather easy. Given the accumulation of very high concentrations of 2-HG in cancer cells compared to alpha-keto glutarate of normal cells, I would imagine the fluorescence character of a wildtype cell versus a secondary glioblastoma would be significantly different. The problem, however, I see with designing a tag for 2-HG isn’t with its similarity to alpha-keto glutarate, however, its with its structure overall. As a whole, the molecule does not have any strikingly unique features and shares functional groups with many molecules that would be in its native environment, so any tag designed to bind to this molecule, I would imagine would also bind to many other molecules, making the data unreliable.

      I do think, however, it would be interesting to genetically screen people to determine if they are predisposed to develop type two glioblastomas. They actually have already determined the most common mutation in IHD1 mutants, which occurs at the second base of codon 132, and causes an arginine to mutate to a histidine (Ohgaki and Kleiheus). I think determining the mechanism behind this base change to see if there is something genetically or environmentally that confers some susceptibility to this base change or a greater chance of this base change happening could be a major progression to reducing secondary glioma rates.

  3. Hi Tommy,
    As I was reading through your summary and the discussion, I noticed that the authors say that the new F18-Gln imaging technique developed here is not “superior” to any of the current imaging techniques that are available. They do not just say this once, but instead repeat it twice, which really emphasizes their point that this new technique is not necessarily better than anything currently available; I found this confusing and counterintuitive, as most authors try to emphasize the significance of their work, rather than downplay it, especially in a journal like Science. Do you have any thoughts about why the authors might have done this? Could it just be that they wanted to ensure that the older F18 imaging methods were not eliminated? On a similar note, these statements really had me questioning the actual applicability of the work—F18-FLT, F18-DOPA, and F18-FET are already used to detect gliomas, and this new method is apparently no better or more accurate. How, then, does F18-Gln actually improve/contribute to the arsenal of testing methods already available to the doctor? I realize that having more information about the tumor can be useful, but is it actually helpful in this case? It just doesn’t seem to share any crucial information that would not be available using another method, and it seems difficult for a doctor to justify adding another test to a cancer patient that is already being subjected to so many tests to confirm their diagnosis and monitor treatment progress.

    1. Hey Michael,

      I would argue that there are a couple reasons the authors showed some restraint when presenting their findings. The main reason, I think, is that there are still some things in the paper the authors need to flesh out before making any drastic conclusions. As you mentioned, the authors addressed a few limitations to their study in the conclusion, so I think any smart reader would have been skeptical had they pronounced that F18-Gln was going to save the world indefinitely. Given the rate that this field seems to be expanding, and the large amount of potential imaging analogs (Jörg-Christian Tonn et. Al), I would imagine the authors rushed to get this paper out to further establish themselves (they had a paper about this in 2011 too) and will most likely follow it up with a paper confirming their findings.

      The main niche I think the authors are trying to occupy, however, is the ability to indiscriminately determine the metabolic status of tumors. While I agree that this field seems to have many PET options, from the information I have gathered, none of these various analogs have been as successful at determining the metabolic status of a tumor before or after treatment as F18-Gln has the potential to be (Jörg-Christian Tonn et. Al). I think this is what the authors mean by using F18-Gln as a complementary tool. in that while other tools might be sufficient to concretely diagnose brain cancer, F18-Gln might prove to be the most effective at determining how far along the cancer has progressed.

  4. Hey Tommy,

    I was curious as to the implications of this study, specifically the possible therapeutic outcomes that could come as a result of this new information. In light of the authors own admission that the efficacy of glutamine imaging techniques already established do not vary wildly from their novel techniques, would it be preferable to instead focus on glutaminase activity with the aim of inhibiting it?

    If one prevailing theme among the literature binds all these analytic techniques together, it appears to be varying levels of efficacy tracking the path of metabolized glutamine into an oncogenic cell, with different techniques better or less suited relative to the location of interest (emphasized here by F18-Gln and Gliomas). In association with targeted glutaminase inhibition within the mitochondria of affected tissues, could these tissue-specific glutamine markers allow a way of inhibiting oncogenic proliferation? In your opinion, do the efficacy of these varied trackers meet an acceptable level to justify such crude techniques as inhibiting glutaminase activity?

    1. Hey Vos,

      I mentioned this above to Michael, I think the main niche the authors are trying to occupy is the ability to delineate between different stages of cancer development, and that their analog seems to show more ability to do this than any analog being used right now (Jörg-Christian Tonn et. Al).

      I’m not exactly sure what you mean by the inhibition of glutaminases, but in terms of fleshing out the mechanics of the authors don’t necessarily track F18-Gln through the cell, but instead just measure its uptake by gliomas. I agree though, that if there were some sort of glutamine analog that could track the catabolism of glutamine through a cancer cell and possibly identify therapeutic targets, that this would be highly beneficial to the field.

  5. Hey Tommy, great analysis, and really great paper choice. It’d be really cool if this eventually developed into a new technique, and I think at that point you could probably take some hipster-cred. I’m going to ask this question because stereochemistry is a big factor in the research I do at Muhlenberg. Do you have any idea why they are using the (2S, 4R) diastereomer? The 2S makes sense, because that’s what makes it “L”-glutamine (even though this is a modified version), but why the 4R? I couldn’t find a specific reason in the paper (was the 4S taken up and consumed by the cancer cells, was it somehow easier to make the 4R than the 4S?) – I’d like to hear your thoughts, partly because I’m intrigued by the stereoselectivity of enzymes.

    Another question, and this is sort of unrelated and I’m sure a big stretch, but I just thought of it and I was wondering if you had any input. Let’s say there’s an anti-gliomal drug developed that’s small and non-toxic to non-gliomal cells. Do you think it would be possible to take advantage of the cancer cells’ rapid uptake of glutamine and bind the drug to glutamine with some kind of leaving group, so that most of the tagged glutamine (with the drug on it) would end up at the glioma (because of the rapid uptake), and then have the drug “fall off” of the glutamine and do its thing? Clearly this is a very idealized scenario, but are there striking issues in it, or have you heard of such a method of drug delivery before?
    Thanks!

    1. Hey Besher,

      As far as stereoselectivity goes, I actually have a clear answer (if you can’t tell, I’m pretty proud of myself for finding this paper). In a previous paper, the authors actually synthesized all four stereoisomers of F18-Gln, and found that the (2S, 4R) isomer was taken up the most by gliomas (Craig B. Thompson et. Al), which is why they decided to use it.

      As far as your second question is concerned, I have no idea if this is viable, but it sounds awesome, so I’m going to do my best to analyze it. As far as glutamine catabolism goes, the first step is to hydrolyze off the side chain amino group to create ammonium and glutamate. I would think that replacing one of the hydrogen atoms on the amino group with an inhibitory molecule might might offer a unique opportunity to get the drug into the cell. As long as this molecule did not interfere with the hydrolysis of the amino group, this might be a cool place to start. Obviously there would be a lot more complications, but you never know, that might be a viable avenue.

  6. Hi Tommy,

    Really interesting paper choice! I enjoyed that this paper contained information on both metabolism and on clinical applications. While reading the conclusion of your summary, I was wondering why the authors claim that this method likely won’t replace the current method of PET scanning with 18-FDG. I understand that this is still a viable method for detecting tumors due to increase glucose uptake, but why wouldn’t this new method replace PET scans in the brain? If the use of 18F-Fgln in PET scans seems to increase the resolution of these images due to less interference, is there still a use for old method? In other words, are there types of brain tumors that do not respond to PET scans using 18-FDG?

    1. Hey Matt,

      I think the confusion in your question stems from what I meant by current methods, and for that I apologize, I’ll be sure to be more specific. When I say that 18F-Gln will not necessarily replace current methods but act alongside them, I meant current methods to signify the use of amino acid analogs, MRI imaging, and things along that nature. I agree with you in that I think the authors concluded that for brain scans, 18F-Gln clearly is more effective at identifying tumors than 18-FDG, however, the other analogs show similar promise to 18F-Gln in some aspects, including tumor diagnosis, so there may be a place in the field for all of these analogs to exist.

  7. Hi Tommy! I really enjoyed this paper! I was excited to see that headway is being made on using PET as a diagnostic tool for areas like the brain that have high rates of baseline glucose metabolism and “light up” even when there is no malignancy. The question I have is in regards to the studies they did comparing glucose and glutamine in inflammatory tissue. What role does inflammation play here? Do you know what the mechanistic or chemical explanation is for the lack of baseline glutamine uptake is as oppose to the high baseline glucose uptake? It is wonderful that the glutamine provides better contrast, I was just wondering specifically how this works. Thanks!

    1. Hey Kelly,

      In terms of inflammation, the authors do not spend a lot of time on the subject, but I think what the authors are trying to prove is that even in the presence of inflammation, 18F-Gln does not give a high background noise. Again though, this section was short and vague, which made it slightly confusing to determine their exact purpose with this experiment.

      Basically, the mechanistic explanation for the different background noises is just that the brain uses a lot of glucose and prefers it as a substrate, so brain cells tend to uptake high concentrations of glucose. The brain, on the other hand, has no increased preference for glutamine, so does not take it up at an abnormally high rate, which reduces background noise.

Comments are closed.