Quantitative Morphology in Kidney Research
February 13-14, 2012 Conference Videos

Panel Discussion: Challenges and Unresolved Issues in Determining Glomerular Number in Autopsy and Biopsy Studies
John Bertram (Chair), Monash University, Australia
Bryan Myers, Stanford University
Julie Ingelfinger, Harvard University
Michael Mauer, University of Minnesota
Kathryn White, Newcastle University, United Kingdom

Video Transcript

00:00:00,000 --> 00:00:08,000
JEFFREY KOPP: And now we’ll have a panel discussion sort of rolling into these same topics. We’ll have John Bertram, Bryan Myers

00:00:08,000 --> 00:00:18,033
unfortunately couldn’t be here, Julie Ingelfinger, Mike Mauer, Kathryn White, and I think some of those individuals may have brought short slide

00:00:18,033 --> 00:00:23,899
presentations, which they’re welcome to show now maybe to launch things.

00:00:23,900 --> 00:00:30,966
JOHN BERTRAM: I’ll introduce our panel, Julie Ingelfinger from Harvard, of course; Mike Mauer from the University of Minnesota; Kathryn White

00:00:30,966 --> 00:00:43,432
from Newcastle University in the U.K.; and Kevin Lemley will also say a few words. Bryan Myers was not able to be with us. So I think you’ve

00:00:43,433 --> 00:00:53,533
heard a fair bit from me and so I think what I will do is ask perhaps the panel to give their thoughts on what they’ve heard this afternoon and their

00:00:53,533 --> 00:01:03,766
views and other concerns or additions or questions or comments and then we’ll open up to the floor. So perhaps, Mike, you don’t have slides

00:01:03,766 --> 00:01:09,866
but you might have some…I’m sure you’ve got some ideas and thoughts you’d like to share with the audience, if you want to do that we’ll just then

00:01:09,866 --> 00:01:18,099
go through the panel and get those ideas out there and then we’ll throw it open to general discussion if that’s all right. So, Mike do you want

00:01:18,100 --> 00:01:28,233
to lead us off?
MIKE MAUER: I think it’s been an excellent set of

00:01:28,233 --> 00:01:43,699
discussions so far and hit some really important points. I might just mention some pitfalls that we need to worry about. John Basgen and I some

00:01:43,700 --> 00:01:58,733
years ago studied dogs to see if we could estimate glomerular number from biopsies and learned some things from doing that. So we did

00:01:58,733 --> 00:02:12,833
MRIs of the dogs and then removed their kidneys, and removing the kidney from the circulation decreased the volume of the kidney by about

00:02:12,833 --> 00:02:31,199
30%, so I think that’s physiologically an important variable. Needles compress tissue and affect the volumes and their relationships with one another,

00:02:31,200 --> 00:02:46,133
and we learned that recently in a study in which one center used the biopsy gun, another center used a Vim-Silverman, and the effect, for

00:02:46,133 --> 00:02:56,999
example, on interstitial fractional volume was about a 15% difference based on the needle alone. So I think it’s very important that, in terms

00:02:57,000 --> 00:03:10,366
of how we sample, that we’re careful that the sampling is very consistent. We looked at glomerular volumes in autopsy material of

00:03:10,366 --> 00:03:28,299
accident victims, compared that to glomerular volumes in biopsy material, and the difference was 100% increase in autopsy compared to

00:03:28,300 --> 00:03:43,533
biopsy of normal individuals that were matched in other ways. So, it’s very important in setting up these studies that they’re very systematic, that

00:03:43,533 --> 00:03:55,333
these variables that contribute to variants are considered in the study design, and it’s very nice to see the work that’s been talked about this

00:03:55,333 --> 00:04:06,833
afternoon because of the consistency and the way the material is handled and the way the studies are done.

00:04:06,833 --> 00:04:13,966
JOHN BERTRAM: All right, thanks very much. You’re completely right and I’m very familiar with the work that you and John did with the combined

00:04:13,966 --> 00:04:23,966
biopsy MR approach. In fact, there’s a recent paper out of Jens Nyengaard’s lab in Denmark doing a similar kind of a strategy and they’re might

00:04:23,966 --> 00:04:32,166
be opportunities there as these imaging—noninvasive imaging—techniques are developed, which I’m sure they will be, to more of that type

00:04:32,166 --> 00:04:40,732
of combination, but the point of how you obtain the tissue in the striking effects on dimensions and ratios and something is probably something I

00:04:40,733 --> 00:04:50,199
wasn’t totally across but it certainly has to be considered very carefully. Kathryn, do you want to get your slides up and perhaps make your

00:04:50,200 --> 00:04:54,900
points please?
KATHRYN WHITE: I actually sort of want to take

00:04:54,900 --> 00:05:04,200
a slight step back as the challenges we’re talking about, I know, are measuring glomerular number. I want to sort of take a step back from the

00:05:04,200 --> 00:05:12,300
methodological challenge and just generally a challenge that I come across is actually convincing people that we need to do it in the first

00:05:12,300 --> 00:05:22,600
place, and so I really want to address a couple of issues—it might be a U.K. thing, maybe we’re lazy, I don’t know—but why aren’t we all

00:05:22,600 --> 00:05:32,066
recounting number and it doesn’t necessarily just apply to glomerular number, it can be cell number and why should it be counted? So I do a lot of

00:05:32,066 --> 00:05:39,632
work with people, with other groups, that have got animal models maybe looking at diseases, and if I suggest that maybe we should count

00:05:39,633 --> 00:05:54,833
glomerular number, podocyte number, or just do stereology. I get the answer it’s too time consuming, it’s too expensive, and yes, it is time

00:05:54,833 --> 00:06:05,699
consuming and time is money so it’s expensive. But if you get meaningful results from your data is, is it too expensive? Is it too time consuming? I

00:06:05,700 --> 00:06:14,200
also get that, “Oh, it’s too mathematical, I don’t understand.” Well I’m not a mathematician. I’ve been doing this now for more years than I care to

00:06:14,200 --> 00:06:22,666
remember. I’m still not a mathematician. There’s a lot of people out there who are mathematicians that are in this room; they’re quite scary, they’re

00:06:22,666 --> 00:06:31,399
very mathematical. You don’t have to be a mathematician to do stereology; you have to be able to count. I can count. You have to

00:06:31,400 --> 00:06:38,933
understand the basic concepts of stereology but you don’t have to design these. There are people out there that are sitting amongst the audience

00:06:38,933 --> 00:06:51,099
here that design these techniques. They’re tools; stereology is a tool. So then I get, “So why bother?” Well, we’ve heard various reasons why

00:06:51,100 --> 00:06:58,166
you should count glomerular number here; there’s a lot of links between low glomerular number in the development of kidney disease and

00:06:58,166 --> 00:07:07,166
hypertension. As I said, I tend to these days look more at animal models of disease, and this can give you an idea of maybe the pathogenesis of

00:07:07,166 --> 00:07:16,099
disease if you look at glomerular number. And I’d just like to show an example here of some work we’ve done on a mouse model of

00:07:16,100 --> 00:07:25,933
glomerulosclerosis which, by six months, has quite severe glomerulosclerosis. Now I get sort of…they say they don’t really want to do

00:07:25,933 --> 00:07:33,966
counting. They’ll sort of send me this up and say, “Can you just have a look at it? Don’t bother with this counting stuff. Just have a look. What do you

00:07:33,966 --> 00:07:48,799
see?” Well, I can see it’s sclerosed and that was at six months. We looked at both in these mice and it’s a WT1 mutation and the wild-type at birth

00:07:48,800 --> 00:07:57,166
have immature glomeruli because nephrogenesis continues for two weeks after birth. So, they have podocytes but they haven’t got foot

00:07:57,166 --> 00:08:07,766
processes. They have endothelial cells but they don’t have fenestration. This is in a lot of the glomeruli. Some are mature but in general you get

00:08:07,766 --> 00:08:17,499
quite a lot of immature glomeruli. In the mutant the majority of the glomeruli were actually more mature; the podocytes had well-formed foot

00:08:17,500 --> 00:08:27,333
processes and the endothelial cells had fenestrations. It was an interesting observation. We also counted glomerular number at one

00:08:27,333 --> 00:08:35,633
month, so this is before there was any evidence of glomerulosclerosis and there was reduction in the number of glomeruli in the mutant compared to

00:08:35,633 --> 00:08:47,166
the wild-type. So, my argument was that yes, the observation was interesting, but if you combine it with robust stereological data you start to be able

00:08:47,166 --> 00:08:55,566
to find out what might be going on; it gives you a lot more information. So now when people sort of say to me, “Do you have to bother with all that,

00:08:55,566 --> 00:09:05,366
can you not just have a look?” Well, I’m a great believer in that what you put in, you get out. So if you do put a bit more of sort of time and effort

00:09:05,366 --> 00:09:13,566
into this, you get more information and hopefully more answers. Sometimes you just get more questions, but hopefully more answers. So yes,

00:09:13,566 --> 00:09:21,499
its time consuming and I agree with John that the developments with MRI are really going to help and we may not have to do such a time

00:09:21,500 --> 00:09:29,233
consuming thing, but I really wanted to say that it is an important thing to be doing.

00:09:29,233 --> 00:09:34,666
JOHN BERTRAM: Ok thanks Kathryn. Julie if you want to follow on please.

00:09:34,666 --> 00:09:49,832
JULIE INGELFINGER: I think my comments will follow on from Kathryn’s and what I need to say is really based on being a clinician who went

00:09:49,833 --> 00:10:02,099
from that to the lab, and so in a number of studies looking a perinatal programming, colleagues of mine and I, at some extent, have looked at

00:10:02,100 --> 00:10:15,033
glomerular number in low protein models in the rat and in high salt models and by counting glomeruli and having people who can do stereology have

00:10:15,033 --> 00:10:30,133
found very important data, and Shao-Ling Zhang who’s a colleague who’s here has been doing this in models of diabetic pregnancy and finding

00:10:30,133 --> 00:10:43,133
more apoptosis and finding differences in numbers of glomeruli. I still look at things as a clinician and so I want to come at these

00:10:43,133 --> 00:10:54,399
comments from the point of view of somebody who is thinking about needs for the future since that’s one of the objects of this conference. What

00:10:54,400 --> 00:11:06,866
I have here is something we all know. If you are unlucky enough to be born early you have a lot of weeks of nephrogenesis remaining and just

00:11:06,866 --> 00:11:19,632
shown here is, by very first grade arithmetic, how many weeks more of nephrogenesis there will be. So we have a very important cohort of

00:11:19,633 --> 00:11:31,633
people whom we could follow longitudinally and about whom we do want to know what happens to glomerular number as well as total nephron

00:11:31,633 --> 00:11:44,166
number because these people get, at least from our survey of our nursery, up to 14 nephrotoxins in the course of their NICU stay; and I’m picking

00:11:44,166 --> 00:11:56,232
this as one example of a cohort of people. So, what we don’t know and what we need is just

00:11:56,233 --> 00:12:04,666
as what I wanted to simply enumerate and that is: we need to find a way to do longitudinal studies with counts or surrogate counts and think that

00:12:04,666 --> 00:12:14,432
would be very important. It may be at first very time-consuming and it will need the kind of study that’s been published that John Bertram

00:12:14,433 --> 00:12:24,999
mentioned today and that will be discussed in detail tomorrow and such better imaging will ultimately bring to the clinic methods to count

00:12:25,000 --> 00:12:36,566
glomeruli in vivo, and what we also need, I think, is a better assessment of kidney function with imaging. What I don’t think we know in the human

00:12:36,566 --> 00:12:48,432
over time is what happens to tubular volume. We know that can certainly increase and decrease in

00:12:48,433 --> 00:12:56,799
a number of things over time and very quickly. We don’t know how much glomeruli change in their volume just in the course of a day. I do know from

00:12:56,800 --> 00:13:08,266
studies I did 30 years ago sieving glomeruli and dumping different vasoactive substances on them, that the volume of glomeruli that we sieved

00:13:08,266 --> 00:13:21,466
would increase four-fold or decrease four-fold very easily, so that would be very important. We need dynamic measures of glomerular filtration

00:13:21,466 --> 00:13:32,999
rate and estimated renal plasma flow with something that allows us to do it serially in the short run to measure something that looks like

00:13:33,000 --> 00:13:44,266
renal functional reserve. The present methods for that are very cumbersome but I think correlating that with glomerular number will be very helpful

00:13:44,266 --> 00:13:59,899
and we need more to look at biomarkers, as Wendy presented today, and we need to find out what that means. So, there’s a lot that we don’t

00:13:59,900 --> 00:14:07,266
know that we need to combine with stereology and being able to count and I’ll stop there.

00:14:07,266 --> 00:14:18,032
JOHN BERTRAM: Thanks very much Julie. Bryan Myers wasn’t able to be with us but Kevin Lemley has volunteered to say and make a few

00:14:18,033 --> 00:14:25,833
comments and then I’ll open it up to the floor for general discussion.

00:14:25,833 --> 00:14:33,499
KEVIN LEMLEY: So, since Bryan couldn’t make it I’m just going to give you, say in Monty Python, now for something completely different: a

00:14:33,500 --> 00:14:40,466
combined functional structural way of estimating glomerular number, because as you saw from John, sometimes we get somewhat disparate

00:14:40,466 --> 00:14:52,132
values. The motivation was we had been engaged for quite a number of years in a study of normal aging and the glomerulus, getting

00:14:52,133 --> 00:15:00,833
intraoperative preimplantation transplant biopsies from living and deceased donors, and we wanted to estimate glomerular number. So our

00:15:00,833 --> 00:15:09,466
surgeon was very nice to us and he would take a wedge biopsy, he didn’t do a needle biopsy, at the preimplantation time and these patients had

00:15:09,466 --> 00:15:16,932
been measured or these donors if they were living had been measured in advance for MRI, so we had a cortical volume, and when we just

00:15:16,933 --> 00:15:25,899
looked at the estimated glomerular density per volume times the cortical volume we got what seemed like very high numbers because we still

00:15:25,900 --> 00:15:37,333
were firm believers in the [---] Benson approach which was 617,000. And so, we tried an alternative method which was to estimate the

00:15:37,333 --> 00:15:48,333
number of functioning glomeruli by measuring with true measures for GFR, single kidney GFR—actually in this case dual kidney GFR divided by

00:15:48,333 --> 00:15:57,333
2—and then divide by an estimate of the single nephron GFR. Now, single nephron GFR can be estimated in humans from the ultrafiltration

00:15:57,333 --> 00:16:07,766
capacity, which we’ll call SNKf, of the glomerulus and the mean transcapillary driving pressure, and in work over quite a few years in the rat we

00:16:07,766 --> 00:16:17,266
found that you can actually get quite good estimates of SNKf from what you can do from an electron micrograph from the surface area for

00:16:17,266 --> 00:16:24,732
filtration and the hydraulic permeability of that area. The driving pressures are from the measured plasma oncotic pressure and an

00:16:24,733 --> 00:16:35,233
estimation of what the intracapillary pressure is; which in rats you can actually measure but not in humans. What we see here are the values we

00:16:35,233 --> 00:16:46,633
got in older and younger donors, because again, we’re looking at an aging effect here, and the biopsies provided an estimation that changed

00:16:46,633 --> 00:16:55,866
quite a bit. Again, if you look at the younger donors they were pretty much in the range that other people were getting and this is, of course,

00:16:55,866 --> 00:17:05,899
the number of functioning glomeruli. When you do it this way it isn’t just what glomeruli are there, it’s the ones that are actually contributing GFR, and

00:17:05,900 --> 00:17:17,266
what was interesting was the measured sclerosis rate could not account for the fairly big decrease with age that you saw in the number of

00:17:17,266 --> 00:17:31,199
functioning, implying that you have a loss of glomeruli over time with “normal aging” that do not leave remnants there. Whoops, that’s for another

00:17:31,200 --> 00:17:35,966
day, I’m not going to show that.
FEMALE: How did you measure the global

00:17:35,966 --> 00:17:40,032
sclerosis rate?
KEVIN LEMLEY: Global sclerosis was measured

00:17:40,033 --> 00:17:46,433
in the biopsy sections with a correction because we estimate the cross sectional diameter because that affects your chance of hitting

00:17:46,433 --> 00:17:55,599
sclerotic, so there’s a small correction because they’re smaller so that you adjust for that.

00:17:55,600 --> 00:18:05,300
JOHN BERTRAM: Okay. So, I’d like to open this discussion up now to the floor if you come to the microscope…the microscope [laughs]…what’s

00:18:05,300 --> 00:18:11,933
that thing called? Yeah. It’s a new kind of microscope. Please say your name.

00:18:11,933 --> 00:18:19,266
JULIE INGELFINGER: I feel right at home at the microscope. That’s a very ingenuous approach to try to look at functioning glomeruli but the

00:18:19,266 --> 00:18:27,032
additional confounder, of course, is the wedge biopsy which I deal with all the time looking at donor biopsies. It’s superficial and not only does it

00:18:27,033 --> 00:18:36,633
overestimate the total number of globally sclerotic glomeruli in the screened donor who doesn’t have overt kidney disease, it also is not

00:18:36,633 --> 00:18:43,866
representative of the single nephron GFR of all nephrons when you make your assessments and your inferences and your assumptions from that,

00:18:43,866 --> 00:18:51,199
so there are several levels of assumptions. We know very well in many other mammalian species that superficial and deep glomeruli don’t have the

00:18:51,200 --> 00:19:00,100
same uniform single nephron GFR.
FEMALE: My name is Marjan Afrouzian from the

00:19:00,100 --> 00:19:09,000
University of Texas. I’m probably one of the few people who have just come to this field to see what beautiful things you are doing and maybe

00:19:09,000 --> 00:19:24,200
it’s a little bit too much to ask but I would like to know if I want to start to come to your field what shall I do regarding my preimplantation biopsies?

00:19:24,200 --> 00:19:37,900
Shall I put them in formaldehyde? What resin method should I do? We need really some guidelines from you for our next patients, not just

00:19:37,900 --> 00:19:49,900
normal kidneys. But this field, whatever you are doing, is going to change what we are doing in pathology for our patients. The way we are

00:19:49,900 --> 00:20:04,866
processing our kidney biopsies or wedge biopsies, for myself I would like to know if you are giving guidelines to us to start coming parallel

00:20:04,866 --> 00:20:12,666
to you with our pathology specimens.
JOHN BERTRAM: Well, I guess I, in my

00:20:12,666 --> 00:20:21,432
introductory talk, kind of put that out there as an idea to try to develop some guidelines and we certainly talked about it briefly as the organizing

00:20:21,433 --> 00:20:32,266
team for the meeting as something that might develop from this. I don’t think any of us underestimate nor any of us believe that would

00:20:32,266 --> 00:20:39,732
be a straightforward and simple or necessarily a quick process, but again, they’ve managed to do it in the lungs. So, if the lung guys can do it then

00:20:39,733 --> 00:20:51,699
maybe those of us south of the diaphragm can have some success as well. I don’t work that much with clinical biopsies so I can’t tell you

00:20:51,700 --> 00:20:58,733
exactly what to do, how to do it now—some people may want to contribute—but that could very well be an outcome from this kind of this

00:20:58,733 --> 00:21:06,766
meeting as a catalyst for it down the track. Jeffrey might have his own thoughts on that particular comment.

00:21:06,766 --> 00:21:14,366
JEFFREY KOPP: No, I have no thoughts, but I’d love to hear Mike or Agnes or other people tackle that one because that is a big question: how to

00:21:14,366 --> 00:21:18,599
apply these stereology principles to clinical tissue.

00:21:18,600 --> 00:21:35,100
MIKE MAUER: I think this a very important question. One approach that one might consider with clinical biopsies would be to initially limit the

00:21:35,100 --> 00:21:53,500
number of sections that are taken and stained for a routine study and let the pathologist decide whether the initial sections are adequate and

00:21:53,500 --> 00:22:05,133
whether they need more and save some of the tissue for studies that you might be able to do more quantitation on. Now this is very much a

00:22:05,133 --> 00:22:18,199
sort of a clinical judgment call of the pathologist who needs to make those decisions but that’s one way in which we might be able, for example,

00:22:18,200 --> 00:22:34,900
that saves tissue for serial sections and for more quantitation, and that will vary in the condition being looked at. So, Agnes obviously is going to

00:22:34,900 --> 00:22:48,966
look at this from the pathologist point of view but let me just give you an example. When we look at FSGS on renal biopsies we sometimes wonder

00:22:48,966 --> 00:23:04,932
why there was such rapid progression when only 25% of the glomeruli showed FSGS lesions. In fact, on serial sections, 25% is about 85% so

00:23:04,933 --> 00:23:19,433
really we are way underestimating it. So, I think it’s at least worth considering our approaches to handling biopsy tissue, clinical biopsy material.

00:23:19,433 --> 00:23:29,599
AGNES FOGO: Thank you for opening the door for discussing renal pathology and diagnosis; that’s a very favorite subject of mine. I think that it

00:23:29,600 --> 00:23:40,566
is very true that we do not have to exhaust the tissue block that is submitted to us for analysis to make a diagnosis but when you the word “limit”

00:23:40,566 --> 00:23:48,299
that could make somebody think that you only have to do maybe one or two or three sections. So, I want to share with you an experience I had

00:23:48,300 --> 00:23:58,933
when I visited China about a dozen years ago and I explained that we routinely did several step levels with several stains at these levels—so in

00:23:58,933 --> 00:24:13,399
our lab we cut 13 slides with a few levels on each slide, so that’s about 40 levels, 50 levels on each biopsy—and I was speaking slowly and this

00:24:13,400 --> 00:24:19,800
was being translated every three or four sentences and I was asked to repeat what I had said, and after I had repeated it three times my

00:24:19,800 --> 00:24:30,300
host in Beijing decided the translator should be fired and they brought in a big board like we have up there and asked me to please draw the slides

00:24:30,300 --> 00:24:40,400
that I stained and sectioned. So, I drew 13 slides, we stained 9 of them, we keep the intervening ones for other studies and drew the tissue and

00:24:40,400 --> 00:24:50,366
the question then came, “But Professor Fogo, you only do this when the diagnosis is FSGS, correct?” I think renal pathologists—my

00:24:50,366 --> 00:24:57,266
colleagues—are really awesome but I don’t think that any of them or I can tell that it’s FSGS before we have done the sections. So the follow up to

00:24:57,266 --> 00:25:05,166
that when I visited China some seven to eight years later, the incidence in Beijing of nephrotic patients with minimal change verses FSGS, the

00:25:05,166 --> 00:25:13,066
ratio had flipped completely from minimal change disease being much more common in biopsied nephrotic steroid-resistant patients to FSGS being

00:25:13,066 --> 00:25:22,866
much more common. So, how many sections can we get? We did the study—maybe you also did—of serially sectioning focal segmental sclerosis

00:25:22,866 --> 00:25:30,766
patients to see how many glomeruli are involved if it indeed is focal; it still is focal when you exhaustedly section. We can section about 80 to

00:25:30,766 --> 00:25:39,499
100 sections—2 micron sections—from a typical core. We don’t do that on the front end. We have been very fortunate to have an ARRA grant

00:25:39,500 --> 00:25:48,900
funded to look at the FSGS clinical trial biopsies. We are getting the unstained tissue sections, the frozen tissue, and what remains in the paraffin

00:25:48,900 --> 00:25:57,733
blocks to do additional immunostaining that we hope will elucidate something about the disease. We can get 10, 20, sometimes 30 sections from

00:25:57,733 --> 00:26:06,999
remaining tissue in the paraffin block. In addition, there are wonderful studies being done by Neptune, Matthias Kretzler, by others, where you

00:26:07,000 --> 00:26:13,700
may on the front end decide that you, with informed consent, are going to do an additional core in the patient and if the diagnosis is clear

00:26:13,700 --> 00:26:23,066
from the initial materials submitted then you will not process and submit that additional core, but all of this comes together. So, I just want to be

00:26:23,066 --> 00:26:31,932
really careful when we talk about limit the use of tissue that you do have to do more than one section and one stain to be able to adequately

00:26:31,933 --> 00:26:38,433
diagnose all of the processes but that doesn’t mean that we have to exhaust the block, and there’s a lot of tissue left in the block. If you have

00:26:38,433 --> 00:26:46,333
excellent histotechs who don’t apply general surgical pathology technique, which is we have to face the block and get well into the tissue so

00:26:46,333 --> 00:26:54,099
we have a good slice right through the middle before we start picking it up, if your histotechs…any of you deal with nephrologists or renal

00:26:54,100 --> 00:27:03,600
pathologists are doing that then you’re throwing away half the tissue which could be used for research or diagnostic purposes.

00:27:03,600 --> 00:27:11,200
MIKE MAUER: I think I’m in total agreement the clinical diagnosis comes first, but I think if we husband the tissue very carefully there may be

00:27:11,200 --> 00:27:22,500
residual tissue in the block that could be looked at in a more careful way in terms of clinical research. One other comment I’d like to make and

00:27:22,500 --> 00:27:36,400
go back to Julie’s point about atubular glomeruli: I think that is an area in which we haven’t looked at sufficiently carefully. What we do know is that

00:27:36,400 --> 00:27:49,933
atubular glomeruli can look like normal but on average they are smaller than glomeruli with proper, normal glomerular tubular junctions, so

00:27:49,933 --> 00:28:00,433
some of the variability that we’re seeing in glomerular volume, that we’re seeing in these high risk populations, may in fact represent

00:28:00,433 --> 00:28:11,999
heterogeneity in regard to the glomerular tubular junction. That could easily be driving the hypertrophy of the residual functioning nephrons

00:28:12,000 --> 00:28:25,766
that are not separated from their tubules and could be an explanation of some of the findings. What is common about atubular glomeruli is that

00:28:25,766 --> 00:28:40,699
very proximal to them are atrial [---] tubules, so if there’s a substantial amount of tubular atrophy in the tissues, that should be something which

00:28:40,700 --> 00:28:46,033
should be suspected.
JEFFREY KOPP: So, I’d like to return to the issue

00:28:46,033 --> 00:28:55,633
of acquiring tissue and make the statement that as nephrologists we used to all use 14-gauge needles and some of us still do, but there has

00:28:55,633 --> 00:29:04,733
been a migration to have a radiologist do it who tend to favor the 16- or the 18-gauge needles that often go into medulla and so don’t yield as

00:29:04,733 --> 00:29:12,499
many glomeruli. So I would say for somebody who’s setting out to try to do some of these stereology techniques, the more tissue the better

00:29:12,500 --> 00:29:20,233
and there are multiple ways to acquire it by a larger needle and several passes and making sure you stay in cortex. But I wanted to come

00:29:20,233 --> 00:29:29,066
back, John, to a couple of more questions about how the autopsy studies were done. You settled on 30 glomeruli, one question is maybe you could

00:29:29,066 --> 00:29:37,066
talk a little bit about that. Presumably you did some prior studies that looked at larger numbers and felt that the parameters tend to destabilize after

00:29:37,066 --> 00:29:42,032
stopping at 30 and then how did you select the 30?

00:29:42,033 --> 00:29:51,133
JOHN BERTRAM: 30 from memory and Wendy can correct me if I’m wrong. We figured we wanted to count the same size, the same number,

00:29:51,133 --> 00:29:59,966
of glomeruli per layer of cortex as Wendy’s already mentioned. So, from memory I think it was a bit of an empirical decision to choose 10 from

00:29:59,966 --> 00:30:11,632
the outer third of the cortex, 10 from the middle third and 10 from the inner third of the cortex. We have looked at some work in the last year or so

00:30:11,633 --> 00:30:18,966
and published a paper in NDT I think about 6

00:30:18,966 --> 00:30:26,666
months or so ago asking the question, “Do we really need to do all those 30?” because it is dissector Cavallieri so it is very time consuming,

00:30:26,666 --> 00:30:39,266
and we found that we could get reasonably robust estimates if we did around 9 or 10 in most glomeruli in most situations. But again, I

00:30:39,266 --> 00:30:46,299
emphasize these are not samples of kidneys from patients with overt renal disease, this is our autopsy series, so in a sitting of pathology that

00:30:46,300 --> 00:30:58,466
number may well be not accurate. I think Kathryn may have done a study some years ago looking at the numbers of glomeruli in biopsies to estimate

00:30:58,466 --> 00:31:03,566
volume with Cavallieri and to get kind of a robust estimate. Do you want to speak to that?

00:31:03,566 --> 00:31:13,066
KATHRYN WHITE: Yeah. We did a study and we looked to see how many glomeruli we had to measure volume in and we did it in 10 per biopsy

00:31:13,066 --> 00:31:24,132
and also compared it—we’d just done 5—and the variation was minimal, so we decided to just count 5 glomeruli per biopsy and it gave us a

00:31:24,133 --> 00:31:33,733
good estimate of glomerular volume. But of course that was in a biopsy core, so you weren’t going to see variations going from one lope to

00:31:33,733 --> 00:31:40,433
another different area of the kidney, so within a biopsy we could get away with just 5.

00:31:40,433 --> 00:31:45,533
JOHN BERTRAM: That was an Epon-imbedded tissue from memory and you were using it in 1-micron sections?

00:31:45,533 --> 00:31:54,566
KATHRYN WHITE: It was, which goes back a bit to this sampling that we’ve been talking about. It sounds like there needs to be a bit of

00:31:54,566 --> 00:32:01,399
communication between pathologists, nephrologists, and researchers but from a technical point-of-view I know that generally

00:32:01,400 --> 00:32:08,300
pathologists seem to still use paraffin-embedded tissue and I don’t know whether—there’s a lot of advances in resin-imbedded tissue that can be

00:32:08,300 --> 00:32:16,000
used for lots of different histological stains and immuno work—whether there’s any sort of room to go to a different embedding medium that may

00:32:16,000 --> 00:32:23,633
be good because you can cut thinner sections, possibly. You get more out of it and I don’t know whether that would be of use between the

00:32:23,633 --> 00:32:29,699
pathologists and researchers. Agnes is shaking her head so she doesn’t….

00:32:29,700 --> 00:32:40,133
AGNES FOGO: You can’t set up a clinical high-quality lab for more time-consuming difficult-to-handle research studies unless it’s funded, so if

00:32:40,133 --> 00:32:55,566
you were going to do that for all clinical biopsies, it would require changes across [---].

00:32:55,566 --> 00:33:06,499
MICHAEL MAUER: Also, it might limit the use of, let’s say, formalin-fixed paraffin-embedded tissue for certain immunohistochemical gene expression

00:33:06,500 --> 00:33:09,666
studies and so on.
AGNES FOGO: [inaudible comment]

00:33:09,666 --> 00:33:19,699
KATHRYN WHITE: There’s lots of different resins than Epon, though. There’s a lot that take histological stains.

00:33:19,700 --> 00:33:26,133
JEFFREY KOPP: So Kathryn, could you enunciate what the advantage of the Epon is? Is it less shrinkage? You talked about the thinner section;

00:33:26,133 --> 00:33:29,099
is that important for a measure of glomerular volume?

00:33:29,100 --> 00:33:35,400
KATHRYN WHITE: It was a study that we were doing, mainly electron microscopy, so it was a resin that we use for electron microscopy and it’s

00:33:35,400 --> 00:33:41,266
not good for histological stains, it will only take one, but there are other resins that you can use but the reason we used that was because we

00:33:41,266 --> 00:33:55,999
predominately do an electron microscopy study.
PAUL KIMMEL: Thank you. I was very struck by

00:33:56,000 --> 00:34:05,233
Kathryn’s and Julie’s comments because I think we also have to step back and take a clinical viewpoint and I am a clinician and I’m sitting here

00:34:05,233 --> 00:34:11,633
and thinking about what can we do with these wonderful data because I know we’re not going to change anybody’s glomerular number; we

00:34:11,633 --> 00:34:21,966
don’t have a good drug for that. So, I would encourage all of us over the next 18 hours that we have, to think about the clinical utility, the

00:34:21,966 --> 00:34:30,532
clinical utility of our measurements. We’re not going to longitudinally biopsy a huge number of patients, although I know that Mike has done that,

00:34:30,533 --> 00:34:42,933
but that’s not going to be a clinical tool. So, I think we have to categorize our studies and the utility of what we have in terms of pathogenetic

00:34:42,933 --> 00:34:54,433
factors, biomarkers I like very much and I can see biomarkers as a potential use, and the clinical applicability, and I think providing a framework for

00:34:54,433 --> 00:35:02,499
what our measurements should be and where our research could go could be very useful for the institute in planning. So, I’d like to encourage

00:35:02,500 --> 00:35:10,400
that kind of structure for where we’re going, what’s research, what’s going to be clinically applicable.

00:35:10,400 --> 00:35:21,700
MICHAEL MAUER: If I could comment...I think that the data that have been presented this afternoon suggest there are some important relationships

00:35:21,700 --> 00:35:38,466
between glomerular number and certain kinds of outcomes or risk. It’s also been suggested that we should spend the money to develop a way of

00:35:38,466 --> 00:35:55,799
estimating glomerular number in living patients. So, the possibility may be that loss of glomerular number is a sensitive indicator of early renal

00:35:55,800 --> 00:36:14,100
disease progression before loss of GFR or proteinuria or other variables come into play, and so it may be important as a risk predictor, it may

00:36:14,100 --> 00:36:26,933
be important as an outcome for clinical trials of early disease. So, it’s taking these laborious methodologies that tell us that glomerular number

00:36:26,933 --> 00:36:37,033
is important, let’s see if there’s a way that we can accurately estimate it in people and then see how it…maybe that is a biomarker.

00:36:37,033 --> 00:36:42,233
PAUL KIMMEL: Well, that’s exactly what I’m talking about, Mike. I think we have to have clinically useful, testable hypotheses.

00:36:42,233 --> 00:36:46,499
So, that’s what I’d like to encourage in the discussions tomorrow,

00:36:46,500 --> 00:36:52,333
which are broken down regarding technical groups. I’m thinking about grants that are going to come in

00:36:52,333 --> 00:36:58,733
and proposals that are going to come in and things that review groups will look at and say, “Well, you know, we’re never going to be able to

00:36:58,733 --> 00:37:07,599
do that,” versus clinically applicable tests that can be done. There’s something nice to do with cohorts. You have to follow these premature

00:37:07,600 --> 00:37:16,666
children for a long period of time if you were going to start at the time they were born. Can you do those studies starting out with 20-year-olds?

00:37:16,666 --> 00:37:26,399
Can you do them cross-sectionally? I’d just like to keep the clinical focus in everybody’s mind because I think that’s going to help us use the

00:37:26,400 --> 00:37:31,600
data better.
FEMALE: One thing I would like to add to what

00:37:31,600 --> 00:37:50,166
you just said is that if there were a way that one could estimate glomerular number at the end of, say, discharge from a NICU—neonatal intensive

00:37:50,166 --> 00:38:05,232
care unit—it would, if there were cohorts, provide a group who could be followed for quite some time. Whether the technology through either

00:38:05,233 --> 00:38:21,133
MR or PET scans with markers will come to pass in the next 5-10 years, I think that will be a very…[inaudible section]…until they’re 20, you can look

00:38:21,133 --> 00:38:24,399
at them much earlier.
PAUL KIMMEL: Right, but then it’s a question of

00:38:24,400 --> 00:38:33,666
when do you assess them? I think that’s what the promise is of non-invasive methods and I think we have to consider them very seriously.

00:38:33,666 --> 00:38:43,332
FEMALE: Could I just support that theme further? I’m sort of an adult nephrologist. I think we need to consider the circumstances under which a

00:38:43,333 --> 00:38:54,166
non-invasive technique for glomerular number and volume and volume heterogeneity—I must get that in—would be useful, and from my point-

00:38:54,166 --> 00:39:02,199
of-view it’s less about the terminal end stages of renal disease than deciding if somebody really has a problem and if they’re progressing from

00:39:02,200 --> 00:39:11,000
early stages to medium sorts of stages, and I think just sort of working out in one’s mind when that might be useful and when it might make a

00:39:11,000 --> 00:39:19,566
difference is important. Then I’d like a discussion or for me to understand: what are the obstacles to developing this technology? I’m of the

00:39:19,566 --> 00:39:30,732
understanding that we have the tools available to us now; that the resolution to pick up a glomerulus is now with us and that there are

00:39:30,733 --> 00:39:40,399
various characteristics of a glomerulus. For example, be it flow or be it water transport or be it a marker or whatever, that it could be identified,

00:39:40,400 --> 00:39:52,366
could be labeled, could be sized. Is it that I don’t understand something about the state of the obstacles to the technique or we have not been

00:39:52,366 --> 00:40:02,266
forceful enough going to our radiological colleagues and imaging colleagues and saying, “Can you work this one through with us? What

00:40:02,266 --> 00:40:11,932
are the approaches?” Now, maybe many of you know what stages have been reached in getting this towards such a technology—non-invasive,

00:40:11,933 --> 00:40:21,866
non-toxic—in humans, non-damaging, that I don’t, but I’d like a discussion about what those obstacles are and what the approaches might be.

00:40:21,866 --> 00:40:30,032
JOHN BERTRAM: We are going to discuss this at some length tomorrow but there’s been a lot of talk today about non-invasive imaging, so maybe I

00:40:30,033 --> 00:40:38,199
could ask Kevin and Norbert Gretz maybe for just one or two minutes each just to give us some things to think about overnight before we discuss

00:40:38,200 --> 00:40:49,433
this at greater length tomorrow because there’s clearly a lot of interest. Norbert, do you want to kick that off, perhaps?

00:40:49,433 --> 00:40:59,933
NORBERT GRETZ: I think the idea is great but what we are currently doing is we are working ex vivo, so we are far off and it’s probably not

00:40:59,933 --> 00:41:11,299
suitable for all disease types so you definitely need an adequate, stable basal membrane. If that is not the case, forget it.

00:41:11,300 --> 00:41:20,233
KEVIN LEMLEY: I want to add to that. We’ve done

00:41:20,233 --> 00:41:31,766
some work in vivo in animals. The problem is, the contrast agent has to be cleared for regulation so we’re working on that at the moment and actually

00:41:31,766 --> 00:41:38,599
moving an agent like that to clinical trials, so we’re using a specific contrast agent that labels the glomeruli; so if you want to do counting, that kind

00:41:38,600 --> 00:41:45,666
of thing. The real place where there needs to be improvement is in sort of moving through the regulatory process for things like that. The

00:41:45,666 --> 00:41:54,132
second part is, the resolution at sort of the in vivo level in humans is not as high as it is in animals specifically because the gradient strength isn’t

00:41:54,133 --> 00:42:03,966
strong enough and I’ll talk about that tomorrow in the meeting, but it’s really just sort of a technological challenge. We have all the tools to

00:42:03,966 --> 00:42:08,132
make it happen, it’s just most people just don’t buy that kind of equipment, and that’s for this specific type of measurement. There are people…Vivian

00:42:08,133 --> 00:42:14,733
Lee who was at NYU and now she’s at Utah, has done a lot of work on contrast agent uptake in the kidneys measuring flow and stuff and I

00:42:14,733 --> 00:42:23,333
think that those things at sort of a lower resolution give you estimates of gs like GFR and that kind of thing. thinJOHN BERTRAM: Bob?

00:42:23,333 --> 00:42:31,066
ROBERT CHEVALIER: Bob Chevalier, University of Virginia. I want to get back to the question of atubular glomeruli that’s been surfacing for the

00:42:31,066 --> 00:42:38,466
last couple of hours. I think the reason that this has been under-appreciated in human tissue is it’s been so difficult to document. Jean Oliver first

00:42:38,466 --> 00:42:46,299
described this with microdissections over 50 years ago and then Marcussen had a series of papers in the 90s showing a large variety of

00:42:46,300 --> 00:42:56,666
diseases using serial sectioning, whether it be renal artery stenosis, transplant rejection, a whole host of disorders characterized by this,

00:42:56,666 --> 00:43:07,266
and it now looks like the rate at which atubular glomeruli develop depends on the disorder. Jess Thoene published a couple of years ago that in

00:43:07,266 --> 00:43:15,232
cystinosis it’s extremely rapid and by the time kids are 10 years old most of their glomeruli are disconnected and it’s not surprising because this

00:43:15,233 --> 00:43:23,966
disease attacks the proximal tubule; and Michael Mauer was Type I diabetes…there’s a paper in Type II that seems to happen later. So, young

00:43:23,966 --> 00:43:38,566
adults get it with Type I, later in Type II. So there’s a spectrum and I do think it’s important to know, when you’re looking at glomeruli, whether they’re

00:43:38,566 --> 00:43:45,566
connected or not and it would seem that the imaging techniques may give us a window in this. The glomeruli are smaller but with the overlap it

00:43:45,566 --> 00:43:53,832
probably is difficult to know what the story is with the connected tubule. We published a paper in the rat ureteral obstruction model in AJP last

00:43:53,833 --> 00:44:04,966
year showing there’s a trick in the mouse that, since there are columnar parietal epithelial cells that are around the urinary pole in a normally

00:44:04,966 --> 00:44:16,432
connected glomerulus, that the Lotus staining goes away if that glomerular tubular junction is damaged. In a survey section it’s possible to get

00:44:16,433 --> 00:44:26,599
the proportion of glomeruli that are staining versus those that are not as a rough measure of the health of the glomerular tubular junction in the

00:44:26,600 --> 00:44:30,900
mouse model.

00:44:30,900 --> 00:44:39,466
MICHAEL MAUER: I think these are important comments and I would throw pyelonephritis into Marcussen’s work showing how important it is in

00:44:39,466 --> 00:44:55,599
that disease. One clue, I think, is to pay attention to Bowman’s capsule. I think that when you start to see towards the tubular pole serious

00:44:55,600 --> 00:45:06,166
abnormalities—re-duplication, thickening of Bowman’s capsule—start to worry about this process. I think looking at the surrounding of the

00:45:06,166 --> 00:45:15,632
glomerulus, if the nearby tubules are atrophic, if the Bowman’s capsule is abnormal, I think your index of suspicion should go up and then you

00:45:15,633 --> 00:45:29,533
want to start looking through the sections and I think we will find many more of these abnormalities than we initially thought. One other

00:45:29,533 --> 00:45:42,933
comment I’d like to make and I think it’s a conceptual one that’s important to keep in mind along with Paul telling us to pay attention to the

00:45:42,933 --> 00:45:57,399
clinical, and that is that we may think differently about the role of some of the variables that we’re thinking about at different stages of disease, and I’ll

00:45:57,400 --> 00:46:04,333
give you an example to illustrate this. If we look at people who are diabetic, who get a brand new kidney in a transplant, and we do research

00:46:04,333 --> 00:46:11,166
biopsies an average of 12 years later and we take people with two kidneys who had diabetes for the same duration as the transplant as being in

00:46:11,166 --> 00:46:32,532
the diabetic environment with similar levels of A1C, etc., then the rate of development of lesions in the one kidney versus the two kidney

00:46:32,533 --> 00:46:45,533
populations is absolutely superimposable. So for genesis, nephron number may not be so critical, at least for diabetic nephropathy. For

00:46:45,533 --> 00:46:57,133
progression, from established advanced injury that turned into uremia is probably a totally different story. So depending on the stage of the

00:46:57,133 --> 00:47:05,533
illness that you’re looking at, the importance of these variables may be very, very different.

00:47:05,533 --> 00:47:09,399
JOHN BERTRAM: I’m cognizant of the time, but perhaps one final question? Thanks.

00:47:09,400 --> 00:47:17,366
TAKAMUNE TAKAHASHI: I am Takamune Takahashi at Vanderbilt University and I have a brief question about the clinical application of the

00:47:17,366 --> 00:47:26,032
glomerular number data. There are many of the obesity people and I think that obesity increased their glomerular volume and

00:47:26,033 --> 00:47:30,066
my question is: does this depend on the glomerular number or is glomerular number independent? So my question is: does obesity

00:47:30,066 --> 00:47:50,799
increase in the glomerular volume depend on the glomerular number or independent?

00:47:50,800 --> 00:47:59,766
JOHN BERTRAM: Certainly, the relationship between BMI and mean glomerular volume is quite clear. I don’t know, Wendy, whether we’ve…I’m

00:47:59,766 --> 00:48:11,732
sure you have looked at the relationship between Nglom and BMI as it applies to the glomerular size. Do you want to comment on that if that was the

00:48:11,733 --> 00:48:16,166
question you were asking?
TAKAMUNE TAKAHASHI: Yes. Do you think that

00:48:16,166 --> 00:48:26,399
obesity affects the glomerular volume? Obesity affects the glomerular volume, do you think?

00:48:26,400 --> 00:48:33,300
JOHN BERTRAM: Yes, I do. In fact, we’ve shown it in terms of average glomerular size and also in the expansion of the variation in the glomerular

00:48:33,300 --> 00:48:39,333
size, absolutely, with a greater spread in those with the larger BMI, for sure.

00:48:39,333 --> 00:48:45,733
TAKAMUNE TAKAHASHI: So my question: is this glomerular number independent or dependent?

00:48:45,733 --> 00:48:53,399
JOHN BERTRAM: I’m not sure on that. Wendy, have we looked at that in terms of the setting of obesity and whether this expansion of Vglom

00:48:53,400 --> 00:49:04,000
relates to the nephron number? I don’t know that we’ve done that.

00:49:04,000 --> 00:49:09,266
WENDY HOY: I’m sorry. I can’t really answer it precisely accurately but they’re both independently significant in multivariate equations

00:49:09,266 --> 00:49:21,266
of influencing Vglom via BMI and lower Nglom and you get a lot of explanation of variants by equations that have them both in, and you get a

00:49:21,266 --> 00:49:30,999
lot of explanation of variants looking at Nglom in the context of current body size. The question is, if people with very marginal low Nglom became

00:49:31,000 --> 00:49:39,766
very fat, would they have more glomeruli hypertrophy than those with robust Nglom? I think the answer is probably conceptually “yes.” I’d

00:49:39,766 --> 00:49:48,966
have to go back and dig around and figure out a way to demonstrate that, but indeed, that’s what the multivariate data suggest.

00:49:48,966 --> 00:49:56,966
JOHN BERTRAM: So conceptually, yes. We may not have exactly done that analysis that you’ve asked for. Okay. Thanks very much and I’ll hand

00:49:56,966 --> 00:50:02,499
over to Jeffrey to wrap things up.
JEFFREY KOPP: Well, thank you to our speakers,

00:50:02,500 --> 00:50:11,066
the discussants, the audience. I think this has been really a remarkable afternoon and early evening discussion, and we’ll continue with more

00:50:11,066 --> 00:50:18,299
tomorrow at 8:00, so I’ll see you then.

Date Last Updated: 10/3/2012

General Inquiries may be addressed to:
Office of Communications and Public Liaison
Building 31, Rm 9A06
31 Center Drive, MSC 2560
Bethesda, MD 20892-2560
Phone: 301.496.3583