Quantitative Morphology in Kidney Research
February 13-14, 2012 Conference Videos
Genetic Variation in Podocyte Number Transcript: Lessons From Mice
Erwin Bottinger, Mount Sinai Hospital School of Medicine
Video Transcript
1
00:00:00,000 --> 00:00:11,066
KEVIN LEMLEY: Our next speaker is Erwin Bottinger from Mount Sinai. He’s Professor of Medicine and he’s Director of the Charles
2
00:00:11,066 --> 00:00:20,299
Bronfman Institute for Personalized Medicine and he’s going to talk about some work I find fascinating, and that’s what determines podocyte
3
00:00:20,300 --> 00:00:29,933
number and I think this is something that’s going to help make sense of what maybe we can measure and maybe we can’t but is a very
4
00:00:29,933 --> 00:00:35,366
important factor we have to keep in consideration. Erwin?
5
00:00:35,366 --> 00:00:50,499
ERWIN BOTTINGER: Wow. Good morning. I would like to thank Kevin for inviting me to talk to
6
00:00:50,500 --> 00:00:56,666
you and I feel profoundly humble that will become apparent as I go through the next 20 or 30 minutes with you, apparently simple-minded, and
7
00:00:56,666 --> 00:01:09,999
so therefore I want to disclaimer. What I’m going to tell you about is not intended to describe methods to take to the FDA as surrogate markers
8
00:01:10,000 --> 00:01:21,333
or methods that would allow you to assess the clinical condition in a patient and then actually make recommendations about the management of
9
00:01:21,333 --> 00:01:32,733
the patient. So, we are interested and have been interested for many years in the processes that underlie the degeneration of the kidney and in
10
00:01:32,733 --> 00:01:47,999
particular, the glomerulus, and so therefore we have in many of our studies recited to experimental animal models and have modeled the
11
00:01:48,000 --> 00:01:59,233
process of chronic kidney disease and glomerulosclerosis over time in many of these models with the understanding that the key is
12
00:01:59,233 --> 00:02:07,199
intervening in certain pathways to find out the relevance of these pathways and the progression of disease and therefore, hopefully,
13
00:02:07,200 --> 00:02:21,233
to identify potentially new targets to prevent or reverse the progression of the disease. So with this preamble let me get straight to the starting
14
00:02:21,233 --> 00:02:33,166
point when I began to think about some approaches to modeling the fate, I shouldn’t say fate, but modeling podocyte number in our
15
00:02:33,166 --> 00:02:47,166
experimental animal models. I was struck by these repeatedly appearing cross-sectional cartoon renderings of an artist in Wilhelm Kriz’s
16
00:02:47,166 --> 00:03:05,132
group where clearly about 15 years ago or so Wilhelm Kriz impressed many of us with a model of the progression of glomerulosclerosis where,
17
00:03:05,133 --> 00:03:21,199
based on the study of many experimental model systems and very careful morphometry, he basically condensed his view of the progression
18
00:03:21,200 --> 00:03:33,800
of glomerulosclerosis to a two-dimensional view and a cross-sectional view and showed, again here as a cartoon, but showed a progressive
19
00:03:33,800 --> 00:03:49,100
loss of podocytes and from largely the periphery, as Kevin just mentioned of the glomerular tuft and where the number of podocytes declined past a
20
00:03:49,100 --> 00:03:57,700
certain threshold, which I’m sure we will hear subsequent talk by Roger Wiggins, can be determined experimentally. So once the decline is
21
00:03:57,700 --> 00:04:10,166
going beyond a certain threshold there will be collapse of the underlying capillary unit and sclerotic process with ensue. So about 12 years
22
00:04:10,166 --> 00:04:22,366
ago or so when I first read these papers I was impressed with this particular process and the rendering of the process and I figured well, you
23
00:04:22,366 --> 00:04:35,166
know, animal models…can we in a very efficient way apply again such a cross-sectional analysis to understand in those models where this has not
24
00:04:35,166 --> 00:04:46,732
been looked at, at the time? What happens to the podocyte numbers? So we applied this to the TGF-beta of one transgenic animal model that
25
00:04:46,733 --> 00:04:56,933
was initially generated in Snorri Thorgeirsson’s lab at the NCI and then extensively characterized by Jeffrey Kopp who is here and his group at
26
00:04:56,933 --> 00:05:11,499
NIDDK, and what Mario Schiffer and Markus Bitzer in my lab could show in a very crude cross-sectional approach staining sections of
27
00:05:11,500 --> 00:05:28,166
kidneys from these animals with WT1 that indeed the number of WT1 positive nuclei per cross-section was declining compared to wild type and
28
00:05:28,166 --> 00:05:38,666
the progressively declining coincident with disease progression. The gray bar here is that, in this context, we can assess the appearance of
29
00:05:38,666 --> 00:05:49,499
the molecular marker, in this case Smad7, in podocytes and you can see that in the wild-type healthy condition a small fraction of podocytes
30
00:05:49,500 --> 00:06:01,000
show the positive signal for a marker of, in this case, TGF-beta-induced apoptosis. But as this disease progressed, virtually all podocytes
31
00:06:01,000 --> 00:06:13,566
became positive for marker of a pathway involved in apoptosis and so therefore giving us clues as to what is going on in a signaling world
32
00:06:13,566 --> 00:06:25,132
in these cells. So, these kind of relatively crude and initial studies were expanded to a number of other models including the Cd2ap knockout mouse
33
00:06:25,133 --> 00:06:38,733
model, which is another model of progressive glomerulosclerosis, not segmental I should argue but rather a global sclerosis and again here you
34
00:06:38,733 --> 00:06:55,999
can see labeling for WT1 in nuclei of podocytes and then counting a large number of glomerular profiles in the large number of animals we could
35
00:06:56,000 --> 00:07:06,333
again demonstrate that as disease progresses there is a loss of podocyte compared to controls and the important aspect is that here the time
36
00:07:06,333 --> 00:07:19,566
factor, as in all models that we’ve reported, is showing that this is a dynamic process because in the same genotype Cd2ap knockout at a pre-
37
00:07:19,566 --> 00:07:28,966
-albuminuric stage the number of podocytes is virtually identical or indistinguishable from the wild-type controls, and only after onset of
38
00:07:28,966 --> 00:07:45,499
proteinuria do we actually see…[phone rings]…Oh, that’s my dean calling. He says, “What are you doing there?” No; just a little joke. This is
39
00:07:45,500 --> 00:07:58,066
clearly observed very consistently in a number of animal models. Others include the [---] podocyte-specific knockout, TGF-B receptor 1,
40
00:07:58,066 --> 00:08:14,966
overexpression in podocytes, etc., etc. Let me move on now to observations that were reported again and have been cited here already, also
41
00:08:14,966 --> 00:08:28,932
over a decade ago by people—some of them are here in the room—that relate to diabetic glomerular injury and clearly established in a
42
00:08:28,933 --> 00:08:42,233
clinical cohort setting that depletion of podocytes is, in fact, a rather early feature in diabetics that develop subsequent albuminuria and progress to
43
00:08:42,233 --> 00:08:53,499
glomerular diabetic injury. So are you all familiar with these studies, at least some of them, the earlier ones, and this applies to both Type I as
44
00:08:53,500 --> 00:09:06,433
well as Type II diabetes; Mike Steffes in Type I and then Pagtalunan and Meyer in the Type II context. Now, at the time I was participating in
45
00:09:06,433 --> 00:09:16,699
the NIDDK-funded Animal Models of Diabetic Complications Consortium and what Tom Coffman and Matt Breyer, as other members in the
46
00:09:16,700 --> 00:09:27,533
consortium, had shown is that certain features of diabetic glomerular injury in mouse models are highly dependent on the genetic context. That is
47
00:09:27,533 --> 00:09:40,366
in two papers that I’m listing here. It was shown that, in particular, the DBA/2J inbred strain was quite susceptible in terms of development of
48
00:09:40,366 --> 00:09:51,299
albuminuria as well as mesangial expansion and sclerosis to experimentally induced diabetes, whereas the black 6c57black/6 inbred
49
00:09:51,300 --> 00:10:06,366
background proved rather resistant. At the time these papers had looked at albuminuria, matrix score, etc., etc., but had not done the careful
50
00:10:06,366 --> 00:10:21,999
analyses of cell numbers, so we in our group hypothesized perhaps the increase susceptibility in the DBA/2 mice is related to an inherent
51
00:10:22,000 --> 00:10:31,566
difference in the podocyte number, and so the hypothesis at the time was that perhaps in the inbred strain that’s susceptible, they start out
52
00:10:31,566 --> 00:10:46,566
with a lower podocyte number per glomerulus and so the environmental injury imposed by diabetes then moves to a threshold of podocyte
53
00:10:46,566 --> 00:10:56,266
insufficiency and point-of-no-return more quickly and rapidly. So this was the underlying hypothesis and we understood that we needed
54
00:10:56,266 --> 00:11:09,832
to be more precise in our analysis as compared to the initial studies, and we came up with a revised protocol that is fully automated, that is
55
00:11:09,833 --> 00:11:20,399
microscopy-based acquisition of glomerular profile images from fluorescence-labeled sections that is then coupled with a software
56
00:11:20,400 --> 00:11:33,300
implemented in the MetaMorph Suite that determines the number of WT1 positive and WT1 negative cells per glomerular sections. So, here
57
00:11:33,300 --> 00:11:45,266
what we do is WT1 label as well as synaptopodin label and DAPI label; so DAPI for all nuclei, WT1 podocytes and then synaptopodin
58
00:11:45,266 --> 00:11:59,099
gives us an estimate of the shape and outline of the glomerular tuft. So, then the software is programmed that when there’s a steep drop in
59
00:11:59,100 --> 00:12:10,200
the synaptopodin signal plus adding a few microns, that will determine again in an automated way the shape and outline of the tuft that is then
60
00:12:10,200 --> 00:12:25,033
also used as the surface area for the tuft and then the cells are counted within that outline. So, here is a fake color-coding. You see the
61
00:12:25,033 --> 00:12:33,999
podocytes and then when DAPI again is blue on the fluorescence but in the software converts it to fake color coding to red, so you can determine
62
00:12:34,000 --> 00:12:48,866
quite readily the nuclei of glomerular cells and count them and also you can impose parameters with regard to nuclear time dimension, etc., which
63
00:12:48,866 --> 00:13:03,799
we did. And so, based on these kind of tuning of this method we developed, I think, a fairly robust, absolutely automated and efficient method of
64
00:13:03,800 --> 00:13:17,100
estimating glomerular cell numbers WT1-negative and WT1-positive cells. This was validated by manual counting and showed a rather decent
65
00:13:17,100 --> 00:13:28,266
consistency between the automated method as well as the manual counting. Now, here came the big surprise. We felt all good that we would be
66
00:13:28,266 --> 00:13:38,999
able to show that DBA/2 susceptible strain would have lower podocyte numbers, but in fact we observed just the opposite. I should refrain from
67
00:13:39,000 --> 00:13:51,033
calling it podocyte numbers; it’s just an estimate. What we’ve clearly shown is that in the susceptible strain there’s a significantly
68
00:13:51,033 --> 00:14:05,566
increased presence of WT1-positive cells per profile compared to the resistant strain. This was a bit puzzling and we needed to then move on
69
00:14:05,566 --> 00:14:16,299
and understand, first of all, the factors that underlie the differences in podocyte number, and secondly, what else could account for the
70
00:14:16,300 --> 00:14:29,466
increase susceptibility as it is relating to podocytes in increased susceptibility to diabetic glomerular injury. So before I go on, we had at
71
00:14:29,466 --> 00:14:41,299
that time been engaged in an extensive genetic mapping approach using mouse genetic tools with recombinant inbred strains, which I will
72
00:14:41,300 --> 00:14:54,200
describe in a little while, to try to map the genomic regions for the differences in the podocyte numbers, but I have to say that as we did these
73
00:14:54,200 --> 00:15:03,900
experiments and repeated the experiments multiple times, initially having significant signals on a genome scan for loci that control podocyte
74
00:15:03,900 --> 00:15:14,566
numbers, on repeated experiments this was not reproducible. So, that was a disappointment but it compelled us even more to look in a different
75
00:15:14,566 --> 00:15:27,366
direction, and that was perhaps that podocytes and the DBA/2 strain more susceptible to diabetic injury compared to the resistant podocytes in the
76
00:15:27,366 --> 00:15:38,032
resistant BL/6 strain. We performed the very initial pilot experiment, Kremena Star at the time in my lab, where she imposed STZ-induced Type I
77
00:15:38,033 --> 00:15:51,666
diabetes for six weeks on inbred BL/6 and DBA/2 animals, and as you can see in the controls, the estimate for podocytes per glomerular section
78
00:15:51,666 --> 00:16:04,999
was increased compared to wild type BL/6 as shown before and in BL/6, indeed, there was very little—at least no significant—change in
79
00:16:05,000 --> 00:16:17,166
WT1-positive cells, whereas in the DBA/2 there was a profound and highly significant loss of WT1-positive cells. So, that led us then to the
80
00:16:17,166 --> 00:16:30,966
hypothesis that there are factors that we can perhaps approach with a genetic mapping strategy that would allow us to understand
81
00:16:30,966 --> 00:16:41,799
podocyte depletion as a factor in the development of diabetic glomerular injury. And Haiying Qi, another graduate student in the lab, picked up on
82
00:16:41,800 --> 00:16:51,000
these initial observations and made her thesis out of this, first confirming what Coffman and Breyer had previously shown, that in fact when we look
83
00:16:51,000 --> 00:17:02,400
at albuminuria in two models of Type I diabetes STZ as well as AKITA genetically driven model of Type I diabetes over time—here are controls, 3
84
00:17:02,400 --> 00:17:14,600
weeks, 6 weeks, 12 weeks of diabetes—this compared to resistant BL/6 very significant and rather impressive increase in albuminuria in the
85
00:17:14,600 --> 00:17:25,900
susceptible D2 strain. Indeed, when we then applied our method of estimating podocytes or glomerular cell numbers, we observed that the
86
00:17:25,900 --> 00:17:38,100
increased albuminuria susceptibility was also associated with the susceptibility to lose podocytes over time in Type I diabetes and
87
00:17:38,100 --> 00:17:49,000
already at three weeks of age we could see significant drop in the WT1-positive podocytes. Interestingly, this reached a level of up to 31% at
88
00:17:49,000 --> 00:18:00,000
12 weeks and I’m sure we will hear some more from Roger Wiggins later on thresholds that would then compel a glomerulus to go down to
89
00:18:00,000 --> 00:18:10,633
complete sclerosis, but 31% loss of podocytes here at a histopathological level was not consistent or was not associated with a sclerotic
90
00:18:10,633 --> 00:18:23,866
appearance, I should point out. So having this tool, having established this basic difference in susceptibility to podocyte depletion induced by
91
00:18:23,866 --> 00:18:35,866
diabetes in these two strains, we recited again to a very powerful mouse genetic model which is the BXD recombinant inbred strain, and just to
92
00:18:35,866 --> 00:18:46,266
introduce you very briefly to the model system, from the parental strains BL/6 and DBA/2 through intercrossing over many generations suddenly an
93
00:18:46,266 --> 00:19:00,932
axis of 20 generations in the crossing of sib links or sib pairs in each generation, one essentially creates a fixed inbred genome in these strains
94
00:19:00,933 --> 00:19:13,499
derived from these initial matings where the genome is a mosaic of the initial parental strains in a fixed configuration. When one applies the
95
00:19:13,500 --> 00:19:22,433
markers that distinguish the parental strains, one can determine exactly or precisely which region on which chromosome was derived from the
96
00:19:22,433 --> 00:19:35,333
parental D2 strain or the parental B6 strain and then this goes into the classic approaches to quantitative trait loci mapping; one can then
97
00:19:35,333 --> 00:19:49,333
assess which of these regions segregate with phenotypic parameters. I should point out that a terrific tool just for reference is the WebQTL
98
00:19:49,333 --> 00:19:57,833
system developed by Rob Williams and colleagues, and I refer you to Google this and take a look if you’re interested in this type of
99
00:19:57,833 --> 00:20:15,066
approach. So together with Matt Breyer and Rob Williams, we set out to undertake an ambitious experiment where we subjected 70 BXD strains
100
00:20:15,066 --> 00:20:27,466
and 5 animals in each strain and the parental strains—you can do the math, it’s quite extensive in terms of animal work—to a long-term diabetes
101
00:20:27,466 --> 00:20:39,632
protocol where STZ-induced Type I diabetes was sustained for 6 months in each animal, and then animals were sacrificed under perfusion protocol
102
00:20:39,633 --> 00:20:54,999
and the intent for our group was to apply our methods to quantitate podocytes as well as take a look at the glomerular tuft surface area after
103
00:20:55,000 --> 00:21:04,800
long-term diabetes. So here you can see, just in the parental strains that were examined in the protocol, we certainly reproduced our initial
104
00:21:04,800 --> 00:21:18,433
findings that, after long-term diabetes, there’s about a 30-40% loss of or reduction in WT1-positive cells in the D2 strain whereas in the B6
105
00:21:18,433 --> 00:21:27,633
strain, even after long-term diabetes, there’s very little change in the WT1-positive cell number. Interestingly, this was associated with an
106
00:21:27,633 --> 00:21:43,066
increase in the tuft area in the susceptible D2 animals. Now, this is a quite illustrative cartoon of the actual phenotype in each one of these
107
00:21:43,066 --> 00:21:50,932
strains, so here you have the parental strains—Black/6, DBA/2—and then here you have the various recombinant inbred strains, and you can
108
00:21:50,933 --> 00:22:04,833
see as you would expect that the phenotype of podocyte number of podocyte per glomerular section tracks or segregate, is quite distinct, as
109
00:22:04,833 --> 00:22:18,266
you would expect, for a genetic model with the two across the strains. So in other words, in most strains or in some strains it’s close to the
110
00:22:18,266 --> 00:22:29,099
Black/6 and in other stains close to the DBA/2 line, and so that is an indication that, indeed, genetics underlie the loss of podocytes in long-
111
00:22:29,100 --> 00:22:43,200
term diabetes. With the computational tools available on the WebQTL online system we were indeed able to map two QTLs, rather narrow
112
00:22:43,200 --> 00:22:55,700
regions on chromosome 13 as well as chromosome 17 in the mouse. You can see the two statistical peaks of the association of
113
00:22:55,700 --> 00:23:08,133
markers with the phenotype and so this discriminates here that these markers point to regions of the D2 genome underlying the
114
00:23:08,133 --> 00:23:21,399
phenotype and is statistically enriched in animals with susceptibility to podocyte depletion. So, here is the location and the locus markers. So, there
115
00:23:21,400 --> 00:23:37,300
are two loci with about 20 genes mapping to each one of those loci. Interestingly, when we applied the same analysis with the change in
116
00:23:37,300 --> 00:23:50,766
glomerular surface area, again based on a rather crude estimate, we find an identical overlapping peak statistically significant on chromosome 13.
117
00:23:50,766 --> 00:24:01,666
So, it appears that one locus on chromosome 13 underlies both expansion of glomerular tuft area as well as loss of podocytes, whereas the locus
118
00:24:01,666 --> 00:24:13,366
on chromosome 17 underlies specifically the loss of podocytes in this animal model system. So to illustrate that even further, we can take a close
119
00:24:13,366 --> 00:24:26,399
up view of the chromosome 17 locus and you see here—although I apologize, this comes out quite small—that here somewhere these are
120
00:24:26,400 --> 00:24:39,800
ranked by phenotype, so from very low podocyte numbers in the 7-8-9 range here to rather quite normal podocyte numbers in the 12-13 range
121
00:24:39,800 --> 00:24:53,300
here you can see that here in green depicts the DBA genome, red depicts the Black/6 genome; you get a very good segregation of DBA genome
122
00:24:53,300 --> 00:25:14,533
with low podocyte numbers at this locus. So, suggesting some dependency of glomerular section area expansion and loss of podocyte
123
00:25:14,533 --> 00:25:28,199
numbers, we see that around the susceptible DBA/2 strain, parental strain value for podocyte number or podocyte estimate as well as surface
124
00:25:28,200 --> 00:25:42,866
area, there’s a cluster of susceptible animals all with the DBA/2 genome content on this locus, whereas in the Black/6 resistant strain there’s a
125
00:25:42,866 --> 00:26:00,866
cluster of animals that have the Black/6 genome at this locus. So, we are in the process of characterizing the genes that are mapping to
126
00:26:00,866 --> 00:26:12,532
these loci. Up here you see there’s, again, about 20 genes. There are some that prioritized to underlie this phenotype that have to do with
127
00:26:12,533 --> 00:26:23,699
control of oxidative stress or generation of oxidative stress, but we really don’t have any conclusive experimental data yet to present. In
128
00:26:23,700 --> 00:26:35,666
keeping with the theme of this meeting, what I want to summarize and leave you with here is that methods that apply nuclear labeling of WT1
129
00:26:35,666 --> 00:26:45,899
as a podocyte marker on a two-dimensional cross-sectional glomerulus section prove, when done in sufficient numbers—and I’ll give you the
130
00:26:45,900 --> 00:26:54,900
numbers below—as robust and enable implementation of high volume experimental protocols that are designed to compare baseline
131
00:26:54,900 --> 00:27:03,500
and change in cell numbers between experimental groups, specifically dependent on intervention or genetic backgrounds and certainly
132
00:27:03,500 --> 00:27:18,033
when observed over time. In our case we find it typically sufficient for statistical analysis to have 50 glomerular profiles in each animal, in 5-10
133
00:27:18,033 --> 00:27:26,833
animals per experimental group, in particular when you are interested in change over time. And so with the method that we have applied,
134
00:27:26,833 --> 00:27:36,733
we can determine WT1-positive, WT1-negative cells per glomerular section area and give a rough estimate of the glomerular section area in
135
00:27:36,733 --> 00:27:54,699
terms of micron-square. I’ve shown you a number of the biological results that we obtained from the application of this approach and we
136
00:27:54,700 --> 00:28:07,100
have identified loci and are in the process of identifying the underlying genes that are related to phenotypic traits in this experimental system. I
137
00:28:07,100 --> 00:28:17,033
want to close with acknowledging people who have done the work, former lab members on the earlier reports were: Mario Schiffer and Markus
138
00:28:17,033 --> 00:28:27,933
Bitzer, and the recent work was done that I described to you in large part by Haiying Qi, a graduate student who graduated successfully
139
00:28:27,933 --> 00:28:40,633
and with honors last fall. Thanks for your attention.
140
00:28:40,633 --> 00:28:52,199
JOHN BASGEN: John Basgen, Charles Drew University, Los Angeles. Well, I think as you heard I wouldn’t report the results as podocytes but
141
00:28:52,200 --> 00:29:02,966
number of podocyte profiles per glomerular cross-section area. But my question is or comment, in your early slides where you showed
142
00:29:02,966 --> 00:29:11,399
the fluorescent immunolabel-like staining and then made the comment that you were surprised that there wasn’t a decrease in the number of
143
00:29:11,400 --> 00:29:24,666
podocyte profiles in the knock-out or whatever—the experimental animals versus the control—and I want to remind people of what John Bertram
144
00:29:24,666 --> 00:29:34,199
said yesterday, that when you’re counting profiles and single sections, the chances of hitting big things are greater than little things and
145
00:29:34,200 --> 00:29:45,333
I’m wondering if the experimental animal was larger in what you were trying to stain and that was the reason why you were seeing
146
00:29:45,333 --> 00:29:53,833
surprisingly more of them and not that there really were more of them if they were bigger.
147
00:29:53,833 --> 00:30:05,499
ERWIN BOTTINGER: Thanks for bringing this up. I think those are good points. Clarification in the particular animal model is [---] knockout model,
148
00:30:05,500 --> 00:30:19,100
what I was referring to was that a point in life of the knockout animals pre-onset albuminuria the appearance of podocyte profiles, if we can
149
00:30:19,100 --> 00:30:27,533
agree on this term, per glomerular section was indistinguishable between the wild-type age-matched mice, as well as the knock-out mice, and
150
00:30:27,533 --> 00:30:38,199
only with onset of the process that is indicative of glomerular injury that is reflected by albuminuria did we actually see a change in the
151
00:30:38,200 --> 00:30:45,600
knockout, but there was no change in the wild-type. So the time component and the dynamic component in these experimental model systems,
152
00:30:45,600 --> 00:31:01,333
that’s what I tried to point out with regard to the volume and particular characteristics of each podocyte profile in the calibration of the method
153
00:31:01,333 --> 00:31:15,833
subsequently, then extensive analysis looking at the distribution of nuclear diameter in various models and we found a certain range of the
154
00:31:15,833 --> 00:31:29,533
nuclear diameter that covered most nuclei on the cross-section, then this diameter was specified; there was a lower diameter and then an upper
155
00:31:29,533 --> 00:31:40,266
diameter specified in the software that then automatically analyzed all sections. So, there was a specified range of nuclear size.
156
00:31:40,266 --> 00:31:44,266
JOHN BASGEN: But that data only came from two-dimensional sections.
157
00:31:44,266 --> 00:31:47,566
ERWIN BOTTINGER: That’s correct. JOHN BASGEN: That, I think, can lead to
158
00:31:47,566 --> 00:31:50,466
misinterpretations. ERWIN BOTTINGER: That’s correct.
159
00:31:50,466 --> 00:32:02,232
ROGER WIGGINS: Roger Wiggins, Ann Arbor. Really, the same question. So, have you tried to make a calculation making some assumptions
160
00:32:02,233 --> 00:32:14,199
about whether…because the mice with the bigger area had the fewer podocytes, so whether that could be accounted for just by the
161
00:32:14,200 --> 00:32:23,200
change in volume in the glomerular tuft, in other words, the fact that you were counting fewer podocyte nuclei, given the assumptions that
162
00:32:23,200 --> 00:32:34,766
you’ve made, could you actually be looking at increased glomerular tuft volume? Because one of your variables…actually they were both
163
00:32:34,766 --> 00:32:47,199
covered by the reduced podocyte number and increased glomerular tuft size were covered by the same marker and just one explanation would
164
00:32:47,200 --> 00:32:52,066
be that either one or the other is correct. ERWIN BOTTINGER: So the question is whether,
165
00:32:52,066 --> 00:33:05,532
just by virtue of having expanding into capillary surface—capillary and matrix, whatever constitutes that—as well as increasing
166
00:33:05,533 --> 00:33:16,566
non-podocyte cell types, potentially does one decrease the podocyte number? I would argue “not” because when we just take podocyte
167
00:33:16,566 --> 00:33:32,566
number…if you argue that increasing size of the glomerular tuft would reduce podocyte numbers then you should actually see a decrease…
168
00:33:32,566 --> 00:33:37,632
ROGER WIGGINS: It would reduce measured podocyte numbers, counted podocyte numbers.
169
00:33:37,633 --> 00:33:47,699
ERWIN BOTTINGER: Yeah, counted podocyte number. Then you should actually see a decrease in the podocyte number with increasing size, but
170
00:33:47,700 --> 00:34:00,466
when we look at the parental strains that is not the case, right? When we look at the parental strains the increased size of the glomerular tuft is
171
00:34:00,466 --> 00:34:08,066
associated with larger numbers of podocytes. Am I not making that clear?
172
00:34:08,066 --> 00:34:10,632
ROGER WIGGINS: Yeah, all right. ERWIN BOTTINGER: So, the starting point is that
173
00:34:10,633 --> 00:34:24,466
you have more podocyte profiles and a somewhat larger tuft in the susceptible strain and what happens over time in diabetes is that the
174
00:34:24,466 --> 00:34:39,032
size of the tuft is increasing whereas the number of podocytes over time is decreasing. So, if you explain the decrease in podocyte number just as
175
00:34:39,033 --> 00:34:48,333
a function of increasing tuft size, where do the podocyte nuclei go? They shouldn’t disappear because we are just doing cross-sections, right?
176
00:34:48,333 --> 00:34:53,566
ROGER WIGGINS: But you’re not doing serial cross sections, you’re just doing a single cross-section.
177
00:34:53,566 --> 00:34:58,299
ERWIN BOTTINGER: Right. JOHN BERTRAM: Tough crowd. It is a tough
178
00:34:58,300 --> 00:35:14,833
problem and I think you’re really actually close to a really smart solution. So I guess the questions John Basgen was raising and Roger is raising is
179
00:35:14,833 --> 00:35:25,533
this issue of what you’re counting of podocyte nuclear profiles as distinct from podocyte nuclei. I guess what you don’t know and maybe you want
180
00:35:25,533 --> 00:35:35,033
to know is the number of podocytes per glomerulus and I think I’m like, perhaps, the situation in the biopsies that Kevin talked about
181
00:35:35,033 --> 00:35:46,666
this morning. In these animal studies you can cut sections of any thickness you like. You’ve got lots of ability to get tissue, you can cut thin sections,
182
00:35:46,666 --> 00:35:52,332
thick sections using beautiful fluorescent markers and it’s really impressive that you’re using something like MetaMorph to start to do this
183
00:35:52,333 --> 00:36:02,833
automatically, which is fantastic. I would think you’re close to be able to with thicker sections and you’re making the sections, you’ve got the
184
00:36:02,833 --> 00:36:09,233
immuno working beautifully, you’ve got MetaMorph doing it automatically in 2D, which is fantastic. With thicker sections you could use an
185
00:36:09,233 --> 00:36:20,033
optical fractionator approach that I mentioned briefly yesterday and probably even get your technique to count nuclei in three-dimensional
186
00:36:20,033 --> 00:36:27,533
space even automatically because you’ve got such specific labeling with your WT1 antibody—I think it is to the podocyte nuclei—and then you
187
00:36:27,533 --> 00:36:36,433
could get, even if you couldn’t do it automatically, you could easily count the number per glomerulus and then Basgen and Wiggins and I don’t ask you
188
00:36:36,433 --> 00:36:46,866
these kind of questions and we’d all come to you and ask you to use your technique. So I think you’re nearly there, but at the moment you’re not
189
00:36:46,866 --> 00:36:54,499
actually getting that number of podocytes per glom and I think if you knew that you could do that, all these issues of “did the glom get bigger,
190
00:36:54,500 --> 00:36:59,933
did the glom get smaller, did the density change,” they’re gone. Those issues don’t matter.
191
00:36:59,933 --> 00:37:07,233
ERWIN BOTTINGER: Well, I’m glad I came here and I can pick up those key insights from experts to improve the method. I should point out that as
192
00:37:07,233 --> 00:37:19,499
long as we can apply an improved method that would get closer to the stringency that you guys are applying to more than 10,000 glomerular
193
00:37:19,500 --> 00:37:30,033
profile, which is what the mapping experiment entailed, as long we can apply it to that with a reasonable effort I think I would welcome any of
194
00:37:30,033 --> 00:37:32,666
these… JOHN BERTRAM: I think you’re very close
195
00:37:32,666 --> 00:37:40,699
because the hard part of the labeling and the whole complexity and sophistication of the genetics approach is incredibly powerful. But I
196
00:37:40,700 --> 00:37:48,933
think with not a lot more development you could probably do this really nicely, whether you could do it automatically, you know push a button and
197
00:37:48,933 --> 00:37:57,333
five minutes later it goes “bing” and it says there’s 156 podocytes, I don’t know; that’s our dream. We’re actually trying to do that but I think you’re
198
00:37:57,333 --> 00:38:00,733
further advanced than we are, but I’d be happy to talk to you about that.
199
00:38:00,733 --> 00:38:07,866
ERWIN BOTTINGER: Thank you. PAUL KIMMEL: Very pretty data, Erwin. So, just
200
00:38:07,866 --> 00:38:19,032
to continue the theme of clinic relevance which we’ve been marking for the last day, I just want to know whether the loci you identified were
201
00:38:19,033 --> 00:38:28,666
associated with renal functional aspects in the animals, such as any kind of estimate of the glomerular filtration rate or proteinuria, to be able
202
00:38:28,666 --> 00:38:36,732
to link the podocyte changes to the disease? You’d expect them to be co-segregated.
203
00:38:36,733 --> 00:38:50,333
ERWIN BOTTINGER: Yes. I can say we just finished a pilot study where we used an inhibitor to the top candidate on chromosome 17 and…
204
00:38:50,333 --> 00:39:00,833
well, let me get back to that. First, the top candidate we considered on chromosome 17 was also increased, a gene product was
205
00:39:00,833 --> 00:39:11,266
increased or a modified gene product was increased in the circulation in the susceptible animals. When we use an inhibitor we prevent
206
00:39:11,266 --> 00:39:25,699
albuminuria and loss of podocyte profiles, and it has to do strictly with reactive oxygen species.
207
00:39:25,700 --> 00:39:36,500
PAUL KIMMEL: But does proteinuria in the animals associate with the loci that you showed us or is it significant only for…?
208
00:39:36,500 --> 00:39:37,533
ERWIN BOTTINGER: Yes. PAUL KIMMEL: It does?
209
00:39:37,533 --> 00:39:42,366
ERWIN BOTTINGER: Yes. This is data from…we were not measuring the proteinuria, this was
210
00:39:42,366 --> 00:39:47,266
measured by Matt Breyer. He was taking on that phenotype===.
211
00:39:47,266 --> 00:39:47,899
PAUL KIMMEL: But you can see they’re links to the same loci?
212
00:39:47,900 --> 00:39:48,600
ERWIN BOTTINGER: Yes.
213
00:39:48,600 --> 00:39:51,100
PAUL KIMMEL: Very nice. Thank you.
Date Last Updated: 10/3/2012
|
|
