Diamond Magnetometry for Measuring Free Radicals in Living Cells

Let me first start with introducing, briefly my research lines. My entire group is centered about diamond magnetometry, which is a new tool to convert magnetic resonance imaging into optical signals.

Why is that interesting? Because this technique has great sensitivity and we use this sensitivity in two different kinds of ways in my research team. On one hand, we use the sensitivity to gain spatial resolutions to see something small. On the other hand, we use the sensitivity to measure something that happened fast, to gain temporal resolution.

In this talk, which is focused on our endeavors in cells, I will mainly focus on the left one (on the slide) so on gaining spatial resolution.

First of all, let me introduce the challenge that we are interested in solving. Free radical generation in cells is a hot topic in biology. The reason is that these molecules play a role in all kinds of stress response in cells. Basically you can say, whenever something is wrong in a cell, there is also a response in free radical generation in the cells.

They are also involved in all kinds of process in the healthy cells like for instance, in immune responses or in signaling or in how cells communicate with one another, but they are also involved in all kinds of diseases including cancer, cardiovascular disease or [inaudible 00:03:39], but they are very short-lived and reactive so that means they are [inaudible 00:03:44].

There is also currently no specific probes available for them so if you mention if you want to see, for instance, a protein and you can have an antibody that binds to this protein and then you can visualize it, but that doesn’t really work for radicals.

Finally, there is a few techniques that can tell you something about radicals and they are listed below, but currently it is problematic to both differentiate and map free radicals in real time. While qPCR, the technique on the left here, this is a technology that tells you something about radicals indirectly. You don’t measure the radicals themselves, but you can measure how cells respond to those. You need many, many cells to get information so you’re losing the spatial component entirely. In the middle is clinical MRI which is typically not mentioned in that context because it’s just not good enough in terms of sensitivity. There is something that you can do with fluorescence microscopy so there are specific probes that react with the radicals and then you can visualize them, but typically they are not very specific for radicals so you can see all kinds of other things as well. You can only measure an increase in radical concentration so you can, for instance not measure inhibition or something like that. Then also, these probes are typically changing which is also problematic if you have long term problems that you’re interested in.

Now, I would like to shortly explain to you a few things about diamond magnetometry. We are interested in defects, more specifically nitrogen-vacancy that change the optical properties based on their magnetic surrounding. What do we do? We take fluorescent nano diamonds, which are provided by Adamas Nano and by Olga, and we bring them into cells and then we take these kind of images that you can see here on the bottom right, this is a confocal image and the bright spots are the diamonds. Then we can go and do certain spectroscopy or do a diamond magnetometry on this specific spots where we have the diamonds. What’s interesting about these diamonds is that you can get the magnetic information by optical means. I will come back to it shortly what exactly we do there.

Quickly, what’s a diamond magnetometer? Essentially it’s a confocal microscope. This is the initial drawing, how we planned to make it and it has a few adaptations to it so we use more sensitive detection and we use built-in microwaves and we have the ability to pulse it. Currently, we have six of these in the labs with different specification and this is one of them that we have actually built.

This is, let’s say a simplified version of the physics behind it. I’m not going to go into the exact photo physics of the NV center, but I’ll explain shortly the scheme that we use the most.

What we currently use the most is something called T1 relaxometry and what you use there is that if you shine with a green laser on your diamonds, you end up in the ground state. Regardless of which state you were before, you can pump in the ground state.

Now, you can use this in the following way. You prepared NV centers in the ground state and then you wait a certain amount of time and then you check if it’s still there.

This is what we do on the right here. This is a curve. The curve is without gadolinium and if you add gadolinium, there is an interaction between the NV center spins and the spin from the gadolinium and then this process is fast. The time that this takes is what we call the T1 time so this is the time that gives your read out for an amount of spin noise that is in the surrounding. Since free radicals are essentially electron spins that are swimming around there, this gives you a great spinning noise in a cellular environment that you can measure with this method.

Here is a cell. In this case it was a macrophage so this is one of the very first experiments of this kind that we did. We take a macrophage, we find our particles and then we do something bad to it. In this case, we more or less poison it.

We are adding DMSO, in this case, this is something that is toxic to the cells and you can basically watch the cell dying. If you look at the curve before in blue and then you look at the curve after, then you can see that you have this decay in fluorescence that we have seen before with the gadolinium much faster when there is oxidative stress or radical production present in this cell. In this case, it’s from apoptosis.

This is the basic principal and now I will show you a few nice examples where we have used this. I will mainly show you one paper that we have recently submitted and this is on the aging of yeast cells.

Why do we study aging in yeast cells? This is a very interesting model system because you can very easily differentiate the mothers and the daughters so you actually know which one is old and which one is young. In this case, you see that the baby is on the left so it is smaller and the mother is the one on the right.

What we did here is that we took these young and old cells and that we compared the free radical load. We looked at these T1 values, which gives you a measure for the radical load and if you have a higher radical concentration, T1 time drops quicker so you have a lower T1. A lower T1 here means more radicals.

You can do much more with this. In this paper, we also compared different genetic strains. The strains that we looked at were strains that have certain genetic modifications that impair the metabolism. What we did then is that we measured the T1 times on the different conditions in these different strains. On this slide, you can see some examples of T1 so the wild type, this is the normal strain that you see on the right. We have sod1 and tor1 so these are the genetic modifications and you see that already from the base level of radicals that are present that they are different and that we can differentiate them with this kind of technique.

We also then subjected these cells to all kinds of treatments and compared how they are doing. In the upper part, we are comparing young cells from the different strains and we subjected them to H2O2 so this is something that stresses them out and should lead to radical formation. We know that in the wild type situation, this is what happens and we also see that in the red curves, that the free radical generation goes up when we stress the cells.

Interestingly in the genetic mutants that’s not necessarily the case. The tor1 shows a slightly different response. This indicates that these genetic mutants are different when it comes to dealing with stress and that leads to a different radical formation. That was also somewhat expected because that’s what these mutants are known for.

If you look to the graphs that are on the bottom, there you have age cells. Here what we did was that we kept the cells for some time, let them age and then we exposed them to the stress again. The first thing that we can see in the yellow data is that once you let the cells age, they also experience stress. What you can see further is that they also respond somewhat differently if you age them and stress them in addition.

Also, I would like to highlight here the gray points. The gray dots and the gray lines indicate that these are single cell experiments from individual cells from individual particles. We can do two things here. We can look at one single cell and tell, for instance which kind of mutation these cells have or what their metabolism is, but we can also look at the bigger population and then compare the different cells which is what we do with the colored arrow bars.

Here is yet another one of these graphs with a lot of these charts. What we compared here is again young and old cells and what we did here is that we gave the cells an antioxidant. The antioxidant should help the cells combat the free radical, or lower the free radical load and that’s exactly what we are seeing. This is the blue bars, but what you can also conclude from this kind of data is that again the genetic mutants respond differently to these kind of stressors, which is quite interesting. In this case, this is of course mutants that we knew and that we suspected would have this kind of behavior, but you can of course also look at Newton’s way. You don’t have this kind of expectations and I think it’s quite nice that you can differentiate between the metabolic rates of these movements here even on the single cell level.

If you look down on the lower panels with the aged cells. Here we are investigating how the situation changes if the cells age in the presence of an antioxidant. With this helper, what you can see then is that they indeed, also the old cells all decrease the radical load when you are adding the antioxidant so that’s I guess kind of expected, but then you can see also that if you add the antioxidant and then add another stressor, that then again the response is somewhat different and somewhat smaller than if you would have the cells aged without these antioxidants.

You can compare all these kind of things. I also have to say that aging, in this case means 24 hours, but we can technically keep the cells for longer. The longest that we have done so far to keep them alive is something like eight days. The only thing that you have to do is you have to keep following the particle and at some point if you do that for too long, the problem with these cells is that they are growing rapidly then the neighbors outgrow the ones that you’re watching. Technically you can do very long term measurements while compare with the conventional techniques, then you have more like minutes to do your measurements and the you see strong bleaching and you can not follow the same cells again.

We have also done in comparison with data from diamond magnetometry, we have also compared with conventional essays, in that case they’re of course not from a single cell level, but we can confirm that we see similar trends of the big colored error bars so these controls we have done.

Then I would like to show you another example from work that we have done on macrophages and what we had planned here was that we want to target our nano diamonds to specific locations. What we did here is we attached an antibody to our diamond, then we brought the antibody for mitochondria. The mitochondria have a high level of metabolic rate, so we are expecting that if a nanodiamond indeed sits on a mitochondrion then you get higher metabolism.

This is indeed what we see and we can compare in situations. If we compare a case where we have nano diamonds at the random location and if we have nano diamonds on the mitochondria, so the two cases are shown here, then we can see that we have a different response. If we are adding CCCP, that’s a decoupler that’s a chemical that is inhibiting a certain radical production in the mitochondria, then you can see that the particles are on the mitochondria responds to the stress while the particles that are at a different location don’t respond.

We also did some measurements in single mitochondria. What we did there is that we isolated the mitochondria from these cells. We did before the measurements that I showed where we measured mitochondria in the surrounding of the living cells, but you can also isolate them and that’s what we did here. We isolate the mitochondria and then we again attach diamond particles and what you can see then is again that you see different metabolic rates if you treat the mitochondria with the certain chemical. What we have compared here on the bottom is on the one hand, we have the diamond magnetometry results and on the other hand we are comparing with a conventional fluorescent assay in this case. Again, the diamond magnetometry is from signal particle and single mitochondria while the conventional assay is from a big ensemble.

This is another example of the kind of curve that we can create with this technique. What we did here is we have internalized fluorescent nano diamonds into macrophage and we are recording nitric oxide signaling so this is something that the cells use to communicate. What we did here is we had several cells and we follow the cells with time and at some point, we added L-Name and L-Name is an inhibitor of a certain process that is producing this nitric oxide.

What you can see, in the blue curve where we added the L-Name is that after a while if you add the inhibitor, it takes a certain amount of time and then after the timing that is expected, you can see the free radical generation go down. This is a case where we measure inhibition which is something that can not really be achieved with a state of the art at least not easily.

[inaudible 00:24:40] … if you just add water. In that case, not much happens so that’s good. And SB, that is an enzyme that would reduce radicals on the outside of the cell [inaudible 00:25:04] to confirm that the diamond particles are indeed inside the cells and only responding to metabolic changes within the cell. Something that I would point your attention to is the nice time resolution that we have in this kind of curve, so if you look at the time axis on the bottom, this is the time is in minutes so what we did here is something that we call a rolling window method. We take, for each data point, we take 5,000 T1 data points and then we move these 5,000 data points by I believe 100 data point forward and then we get a time point roughly every six seconds.

This is another way to represent this kind of data. What we see here is also one of these particles in the macrophage and what we have here is a 3D trajectory in space. What you can see here are coordinates and the particle is moving throughout the cell. In this case, we are recording T1 time versus location. That’s also something we can do so we can use the natural movement of the particle to map the magnetic field in space.

To sum up what I have shown you, we have achieved uptake into cells [inaudible 00:27:11] the most tricky ones here because those have a thick cell wall. We have control measurements with chemicals. I have not really shown you the calibration measurements, but we have done measurements in controlled environments where we measure chemicals that we can then compare what a certain T1 value means so that we can attach a concentration value to a T1 value that we get.

We have the first magnetometry results from the interior of cells and stressed cells and we are currently looking into all kind of different applications that we can study with this. We have already shown that we can study inhibitors which are difficult for the state of the art. We can differentiate between these knockout strengths so those are these genetically modified cells that have a different metabolism.

We can measure the effect of antioxidants and toxins on aging and we can follow the aging process quite nicely due to this nice property of the nano diamonds that they are so stable and that they can be looked at on the long term. Apart from applying this technique to all kind of different things, we’re also currently working on implementing more complex pulsing sequences which will hopefully give us better differentiation capabilities so that we can not only say how many radicals we have, but also say which radicals exactly we have.

With this, I’m already at the end of the presentation and I have a few people to thank for giving me money. This is all these logos on this slide and of course, my research group which is doing all the work and I want to thank you for your attention and Olga again for the invitation.

Olga:

Question from Rajesh Tamaj. What is the concentration of NV centers in nano diamonds in these measurements.

Romana:

What we are using are 70 nanometer nano diamonds. They are irradiated. The concentration of NV centers is around 500 NV centers per particle, on average.

Olga:

It’s about a couple PPMs.

Romana:

Yeah.

Olga:

The next question is from Participant. Was a nice talk. Can you explain why the gadolinium ions shortened the T1 time? In addition, considering magnetical ions can affect the T1 time, is it possible that T1 time of NV was affected by some magnetical ions not radicals?

Romana:

First of all to the gadolinium ions, the reason why they lower the T1 time is because they have a strong spin, that’s also why they are used as contrast agent in clinical MRI so it’s the exact same mechanism that they use there and we are measuring the spin noise from the gadolinium ions.

Could magnetic ions change the T1 times? Technically anything that can cause a spin noise can cause a change in T1 time so we are mostly sensitive to things that are paramagnetic. That can, of course have an influence. If you have multiple things happening at the same time, that could be problematic so to some extent we also trigger certain process and then we measure the process so if you have these kind of cases where you are expecting a certain response, then you can assume that this is the response.

Magnetic ions themselves, there would be iron but I wouldn’t be aware of stress response quick changes due to iron so what we very often look at is changes due to a stressor so we measure at a specific location with the same particle then we add the stressor and then we measure how it responds.

I think it would be very unlikely that this also leads to a big generation of magnetic ions or something like that. In cases where this happens, we would have crosstalk, of course.

Olga:

I’m reading the next question from Adash Radidea. Going back to Heisenberg’s uncertainty principle, how does the pulse frequencies affect the generation of free radicals in cells? How does this affect the measurement of the native state of the cell?

Romana:

That’s an interesting question and we of course always compare to control. If you see, for example on one of the slides I had it very much next to each other, when you add water and when you add the stressor. When you add water, basically nothing so in that case we also have the laser on for the exact same times and we do the exact same thing and you see virtually nothing. With these measurements, due to this case you always have to be very careful and compare to the control.

Something I didn’t show, but that we always do is that we always also test what if you add the particle and only add the stressor, like leave the cell out? What if you have the cell but not the stressor? Test all the individual components is what you also have to do always and there is some cases where you see something like a small change in one.

Olga:

And some related principle, not Heisenberg’s principle level, but at some much simpler level. What is the spatial resolution of T1 measurement of free radicals location within the cell? Is it micron, sub-micron levels? A flavor of spatial resolution of different free radical concentrations measurement possibility.

Romana:

Did I understand correctly, you’re asking about the spatial resolution of these measurements? Right?

Olga:

At the distance if you have a gradient of free radicals within the cell itself, what would be your spatial resolution of when you can tell there is a difference in free radical concentration?

Romana:

I would say that this can very much differ. In some cases, the particle moves so quickly that we at some point lose it, so that’s sometimes the trajectory just stops because we lost the particle. Then if the particle starts to move too fast, that limits your resolution.

Romana:

Typically, I would say about the option that you have. What we do is we sit on a particle and then after a certain amount of time, we recenter the particle and we record the trajectory where it has moved so on one of the last slides I showed this 3D trajectory where you see position and that was one where you see the spatial resolution over time. Then it’s on the order of a micron or something like that.

Olga:

I have a few more questions. From Rolf Rader. What about problems with oxygen molecules diradical? Couldn’t that stress not even inference the reductive environment inside cells? What is the result of higher concentrations of oxygen molecules and T1 values?

Romana:

Oxygen is indeed paramagnetic. If there would be drastic changes in the oxygen level, we could also see that. I have to think about it. We have some data at some point. In a previous group we have compared different levels of oxygen and found a comparably small effect on the T1 values, but I don’t really recall now how exactly this quantitatively with what we have here. If I recall correctly, oxygen doesn’t produce for some reason a very strong signal. It’s paramagnetic and anything that is paramagnetic shows up to some extent. Then again, unless it is really caused by the stressor or by the antioxidant or something like that, it would also show up in the control.

Olga:

Next question is from Quishi Guam, could you comment on the sensitivity of this technique in terms of radical concentration in the measurements presented?

So a question on sensitivity in terms of radical concentration what is the level of detection?

Romana:

Yes. We can detect concentrations down to the nanomolar level. We know that because we have done measurements in controlled environments. For the radicals, we for instance have H2O2 we already did with UV light. That’s one way. Or there are certain chemicals you can create radicals by inducing a chemical reaction. We did that to produce a controlled concentration of radicals and then you can by that way, you can then compare with the T1 signal that you have and it is roughly in the nanomolar to micromolar range.

What we have in the cells, it varies a lot and especially locally there is big variations, but it is somewhere between nanomolar and molar ranges.

Olga:

Rafius Aradafuna asks do you have TEM of your nano diamonds in the cell?

Romana:

We don’t have TEM, we did SEM at some point. We made slices and then we made the SEF images of the slices. Not on everything that we do because it is a very cumbersome process, but for some cells we definitely did. We did for the yeast cells and we did for the macrophages, I believe and also for HT29, that’s the colon cancer cell line.

We also have Phipps M so you get the 3D electron microscopy reconstruction.

For a lot of cells, we do not for absolutely everything.

Olga:

Next question is from Graham on calibration. Could you comment on your process for calibrating T1 sensitivity to radical species concentration? What degree of confidence do you have that such calibrations correlate well to the much messier cellular environment where many species are intertwining?

Romana:

Actually, I’m not sure if I have at least partly answered it, I think. What we did there is that we measured chemicals in known concentrations and then we measured T1 times and then we get an error bar for the calibration.

I have to say that we actually get much better reproducibility inside the cell values than we get in solution with the chemicals because it’s actually not so easy to produce a standard concentration of radicals so you have to induce a reaction and then measure the reaction. While in the cell, it’s tightly controlled so actually we are much more accurate in the cell than we are in this chemical case.

If you want, I can send you the draft for the paper to show you the proof of how accurate it is. We essentially made a calibration curve and then you can; It also depends on which range. There’s a concentration range where you get the best accuracy and if you’re above that, you go in saturation and if you’re below that, you’re below detection limits.

Olga:

From Analese Kraff. How do quantitative measurements of radical levels compare with conventional fluorescence dyes? Do they give similar results?

Romana:

First of all, how does it compare to fluorescent dyes? Mostly we get quite good agreement, but I also have to say that you measure something different and it’s not necessarily always the same. There is for instance radical dyes but usually they are less specific or more specific so there’s really nothing that measures the exact same quantity.

Mostly it’s in good agreement so what we compare to is DCFDA so that’s one of the dyes that measures [inaudible 00:46:34] and in the examples that I’ve shown you for the aging cells, we get more or less the same or very similar results there. Due to the lack of data on the use of in patients with severe hepatic impairment (Child-Pugh scale, class C), Tadalafil is prescribed with caution in this group. Daily use of the drug has not been thoroughly investigated in patients with hepatic insufficiency. There’s also a difference that you have to take into account. In the DCFDA case, you look at an ensemble of cells while in this case, it is a single cell. Then if the result is different it’s probably because you did something wrong or the method is different and you measure something different.

In most cases, we get quite good agreement.

Olga:

Is the instrumentation discussed here available commercially?

Romana:

The one we have is home-built equipment. I believe there is also equipment that is available commercially if you have enough money. It depends on the specification that you need. I think that some things are available commercially. If you are still developing a lot, it is much cheaper to build yourself at this point I think.

Participant:

First of all, Romana very nice talk and you can see how much interest that raises. I have a couple of questions. First of all, in your experiments I’m sure you have done many experiments with different nano diamonds detecting radicals and so on. Did you ever observe inter radical activity of any nano diamonds themselves without aiding any other inter radical molecules?

Romana:

You mean that you get radical generation with diamonds themselves?

Participant:

No, I mean you have suppression of radical generation with nano diamonds themselves so they demonstrated inter radical activity. Did you observe anything like that?

Romana:

To some extent we see that in high concentration as well. What we did there is we did our conventional assays, these fluorescent probes and we measured all kinds of diamond particles to test bio-compatibility and to see if we are inducing something. In the higher concentrations, we indeed also see something like that. It’s a small effect –

Participant:

In high concentration?

Romana:

In the concentrations that we typically use, it’s at least below detection limits. I also have to say that we work with very low concentrations because we have to track the particles and if you have multiple particles then you can jump from one particle to the next if they are too close to each other during the measurements so you want to keep the particle concentration low enough that your measurement doesn’t jump from one particle to the other.

Participant:

What are high concentrations which you would start to see this effect? Approximately?

Romana:

What we typically use for the magnetometry is between one and five microgram per mL, depending on the cell type. Some cells take up particles more, some less. I think the concentrations where we saw that was more like hundreds, hundred microgram per mL. We also did measurements at 10, but I think there we didn’t see it yet. I can send you the paper if you like.

Olga:

How much does nanodiamond concentration affect the resolution and sensitivity of your technique? If there is any hints on nanodiamond concentrations that are optimal.

Romana:

Yes, but I think it has something to do with the previous answer. If we use too high concentrations we run into the limit that we can not measure any more because it jumps from one particle to the other. I think we run into that problem first. It’s fairly limited to how high I go with the concentration.

Romana:

We are also trying to not go too high because you then indeed see all kinds of strange effects from the biology so then you can indeed have this effect that I have just discussed with Badam that you have either some toxic effects or some inhibition so you want to use a low concentration so you don’t influence the biology that you want to measure.