WEBVTT 00:00:00.000 --> 00:00:02.500 align:middle line:84% JEFFREY MARK DRAZEN: So this is how I view pharmacogenomics. 00:00:02.500 --> 00:00:04.770 align:middle line:84% It's the use of genetic information 00:00:04.770 --> 00:00:08.039 align:middle line:84% to determine who will respond favorably or unfavorably. 00:00:08.039 --> 00:00:10.470 align:middle line:84% And I think, actually, unfavorably 00:00:10.470 --> 00:00:14.440 align:middle line:84% may be a better phenotype to a given type of treatment. 00:00:14.440 --> 00:00:18.460 align:middle line:84% And the reason I say that is that, especially the condition 00:00:18.460 --> 00:00:20.760 align:middle line:84% that I study, asthma, there are a lot 00:00:20.760 --> 00:00:27.220 align:middle line:84% of folks that have a placebo effect from treatment. 00:00:27.220 --> 00:00:29.730 align:middle line:84% And so if I'm studying a drug and I give everybody 00:00:29.730 --> 00:00:30.390 align:middle line:90% a placebo-- 00:00:30.390 --> 00:00:31.290 align:middle line:90% that's my trial. 00:00:31.290 --> 00:00:35.670 align:middle line:84% I compare placebo A to the pink placebo, to the yellow placebo. 00:00:35.670 --> 00:00:38.340 align:middle line:84% And I know this is just a sugar pill 00:00:38.340 --> 00:00:42.070 align:middle line:84% and shouldn't have any biological impact on asthma. 00:00:42.070 --> 00:00:44.610 align:middle line:84% But I know that 25% of people are 00:00:44.610 --> 00:00:48.040 align:middle line:90% going to get better from this. 00:00:48.040 --> 00:00:53.490 align:middle line:84% And if I'm stratifying my outcome by genotypes, 00:00:53.490 --> 00:00:56.190 align:middle line:84% you'll see that I could be misled. 00:00:56.190 --> 00:00:59.200 align:middle line:84% But if I'm looking for a phenotype 00:00:59.200 --> 00:01:02.430 align:middle line:84% which is less likely to be contaminated by a placebo 00:01:02.430 --> 00:01:03.690 align:middle line:90% effect-- 00:01:03.690 --> 00:01:09.588 align:middle line:84% for example, people who get worse with treatment-- 00:01:09.588 --> 00:01:11.880 align:middle line:84% then I'm looking at the natural history of the disease. 00:01:11.880 --> 00:01:16.200 align:middle line:84% And I'm less likely to be confounded by that. 00:01:16.200 --> 00:01:20.400 align:middle line:84% Now if you're studying something like cancer or heart 00:01:20.400 --> 00:01:23.520 align:middle line:84% disease with a hard end point, which is death-- 00:01:23.520 --> 00:01:27.030 align:middle line:84% heart disease with a soft end point, which is chest pain-- 00:01:27.030 --> 00:01:27.780 align:middle line:90% it's hard to know. 00:01:27.780 --> 00:01:29.572 align:middle line:84% But heart disease with the heart end point, 00:01:29.572 --> 00:01:31.450 align:middle line:84% something that's easy to define, you 00:01:31.450 --> 00:01:32.700 align:middle line:90% don't have this big a problem. 00:01:32.700 --> 00:01:35.880 align:middle line:84% And you can begin to use favorable and unfavorable 00:01:35.880 --> 00:01:36.550 align:middle line:90% responses. 00:01:36.550 --> 00:01:40.230 align:middle line:84% So I think it's a little more robust phenotype. 00:01:40.230 --> 00:01:42.150 align:middle line:84% What it means is that whenever I do a study, 00:01:42.150 --> 00:01:45.240 align:middle line:84% I have to think about the contaminating effect of this. 00:01:45.240 --> 00:01:46.620 align:middle line:90% Looking at my data. 00:01:46.620 --> 00:01:48.210 align:middle line:84% I know that some of the people whom 00:01:48.210 --> 00:01:51.960 align:middle line:84% I think are better because of treatment or, in this case, 00:01:51.960 --> 00:01:54.780 align:middle line:84% because of a treatment by genotype interaction 00:01:54.780 --> 00:01:57.855 align:middle line:84% are just better by chance, and that there 00:01:57.855 --> 00:01:59.460 align:middle line:84% are more likely to be people better 00:01:59.460 --> 00:02:02.550 align:middle line:84% by chance in the placebo group than there would 00:02:02.550 --> 00:02:04.828 align:middle line:84% be if there was a no treatment group, just 00:02:04.828 --> 00:02:06.370 align:middle line:84% because of the nature of that effect. 00:02:06.370 --> 00:02:08.340 align:middle line:84% So it's a little bit of a contamination 00:02:08.340 --> 00:02:10.240 align:middle line:90% but not a big problem. 00:02:10.240 --> 00:02:14.130 align:middle line:84% So the idea of pharmacogenetics is that there's variability 00:02:14.130 --> 00:02:15.760 align:middle line:90% in treatment response. 00:02:15.760 --> 00:02:17.820 align:middle line:84% So if I treat a bunch of people-- 00:02:17.820 --> 00:02:20.485 align:middle line:84% that some people are going to get better. 00:02:20.485 --> 00:02:22.110 align:middle line:84% Some people are going to stay the same. 00:02:22.110 --> 00:02:23.915 align:middle line:84% Some people can get a little worse. 00:02:23.915 --> 00:02:26.040 align:middle line:84% Since we don't think the medicine makes them worse, 00:02:26.040 --> 00:02:28.110 align:middle line:84% usually we think that it's the disease getting 00:02:28.110 --> 00:02:32.380 align:middle line:84% worse because the medicine is ineffective. 00:02:32.380 --> 00:02:35.970 align:middle line:84% So the first thing that you want to do 00:02:35.970 --> 00:02:39.120 align:middle line:84% if you want to study the pharmacogenetics of anything, 00:02:39.120 --> 00:02:41.400 align:middle line:84% you want to look to see how variable the treatment 00:02:41.400 --> 00:02:45.550 align:middle line:84% response is among members of a population. 00:02:45.550 --> 00:02:47.880 align:middle line:84% And there needs to be substantial variance 00:02:47.880 --> 00:02:50.230 align:middle line:84% in treatment response among members of a population. 00:02:50.230 --> 00:02:52.980 align:middle line:84% So I'm going to give you an example now, from the asthma 00:02:52.980 --> 00:02:58.290 align:middle line:84% world, of a study in which there was variation in the treatment 00:02:58.290 --> 00:03:00.610 align:middle line:84% response among members of a population. 00:03:00.610 --> 00:03:03.300 align:middle line:90% So this is a clinical trial. 00:03:03.300 --> 00:03:04.920 align:middle line:90% It's easy to understand. 00:03:04.920 --> 00:03:07.230 align:middle line:84% Here are patients that are enrolled 00:03:07.230 --> 00:03:09.690 align:middle line:90% with mild to moderate asthma. 00:03:09.690 --> 00:03:12.600 align:middle line:84% And this is asthma which is more of a nuisance 00:03:12.600 --> 00:03:14.440 align:middle line:90% than a real medical problem. 00:03:14.440 --> 00:03:16.320 align:middle line:84% So if patients are not treated, it 00:03:16.320 --> 00:03:18.340 align:middle line:84% doesn't represent an ethical issue. 00:03:18.340 --> 00:03:21.300 align:middle line:84% So that's how we can get away with a two-week placebo run, 00:03:21.300 --> 00:03:25.710 align:middle line:84% in theory, because it's OK not to treat people. 00:03:25.710 --> 00:03:28.890 align:middle line:84% And are then randomized to receive beclomethasone, which 00:03:28.890 --> 00:03:31.680 align:middle line:84% is an inhaled steroid and is one of the standard asthma 00:03:31.680 --> 00:03:34.770 align:middle line:84% treatments compared to montelukast, which 00:03:34.770 --> 00:03:38.980 align:middle line:84% is an antagonist to the action of the cysteinyl leukotrienes 00:03:38.980 --> 00:03:41.400 align:middle line:90% at the CysLT1 receptor. 00:03:41.400 --> 00:03:43.200 align:middle line:84% And the primary outcome in this and a lot 00:03:43.200 --> 00:03:44.760 align:middle line:84% of the data I'm going to show you 00:03:44.760 --> 00:03:49.630 align:middle line:84% is the forced expiratory volume in one second to 12 weeks. 00:03:49.630 --> 00:03:52.140 align:middle line:84% So as an aside by way of pulmonary physiology, 00:03:52.140 --> 00:03:53.760 align:middle line:84% I don't think we recognize any of you 00:03:53.760 --> 00:03:58.320 align:middle line:90% from suggestive pulmonary class. 00:03:58.320 --> 00:04:00.750 align:middle line:90% No, none of you are taking that. 00:04:00.750 --> 00:04:03.002 align:middle line:84% If you breathe in all the way to the top 00:04:03.002 --> 00:04:05.460 align:middle line:84% and then breathe out as hard and fast as you possibly can-- 00:04:05.460 --> 00:04:07.335 align:middle line:84% something like you might go on your birthday. 00:04:07.335 --> 00:04:09.390 align:middle line:84% If you have a lot of candles to blow out-- 00:04:09.390 --> 00:04:12.810 align:middle line:84% how fast you can breathe out beyond a certain level 00:04:12.810 --> 00:04:14.850 align:middle line:84% of effort is an intrinsic property 00:04:14.850 --> 00:04:20.100 align:middle line:84% of your lungs and airways and not how hard you're trying. 00:04:20.100 --> 00:04:22.830 align:middle line:84% So you have to try hard enough to reach that plateau. 00:04:22.830 --> 00:04:24.960 align:middle line:84% And that is not a superhuman effort. 00:04:24.960 --> 00:04:27.120 align:middle line:84% Almost everybody with a little bit of effort 00:04:27.120 --> 00:04:32.430 align:middle line:84% can reach that level where how fast they can breathe out 00:04:32.430 --> 00:04:36.270 align:middle line:84% is not related to how hard they're trying to breathe. 00:04:36.270 --> 00:04:39.180 align:middle line:84% So that if we measure the amount of air that comes out in one 00:04:39.180 --> 00:04:42.150 align:middle line:84% second , we're measuring an intrinsic property in the lungs 00:04:42.150 --> 00:04:44.450 align:middle line:84% and not the motivation of the subject. 00:04:44.450 --> 00:04:46.200 align:middle line:84% And it's beyond a certain level of effort. 00:04:46.200 --> 00:04:48.630 align:middle line:84% And the way you determine whether your subjects are 00:04:48.630 --> 00:04:52.100 align:middle line:84% motivated is usually do two or three maneuvers. 00:04:52.100 --> 00:04:53.850 align:middle line:84% We have somebody coaching them through it. 00:04:53.850 --> 00:04:54.660 align:middle line:90% Deep breath in and out. 00:04:54.660 --> 00:04:57.060 align:middle line:84% Blast it out so you get to see whether they're actually 00:04:57.060 --> 00:04:57.990 align:middle line:90% trying. 00:04:57.990 --> 00:05:00.930 align:middle line:84% And if this test is done correctly, 00:05:00.930 --> 00:05:06.620 align:middle line:84% the variability among tests done in a row is 3% or 4%, 00:05:06.620 --> 00:05:09.470 align:middle line:84% even among you all who've never probably done this test before. 00:05:09.470 --> 00:05:11.098 align:middle line:90% Maybe one of you have asthma. 00:05:11.098 --> 00:05:13.640 align:middle line:84% Statistics would be that one or two of you would have asthma, 00:05:13.640 --> 00:05:15.170 align:middle line:90% have done tests like this. 00:05:15.170 --> 00:05:17.960 align:middle line:84% Without any coaching or with minimal coaching, 00:05:17.960 --> 00:05:22.400 align:middle line:84% can get tested and reproducible within 3% or 4% off the street. 00:05:22.400 --> 00:05:26.360 align:middle line:84% So it's a very useful test because it's easy to do. 00:05:26.360 --> 00:05:27.980 align:middle line:84% And we use it as an outcome indicator 00:05:27.980 --> 00:05:29.330 align:middle line:90% in a lot of breathing trials. 00:05:29.330 --> 00:05:31.010 align:middle line:84% And just to help get you straight. 00:05:31.010 --> 00:05:34.400 align:middle line:84% A bigger FEV1, the more air you can get out in one second, 00:05:34.400 --> 00:05:36.210 align:middle line:90% the better off are. 00:05:36.210 --> 00:05:39.140 align:middle line:84% So that's kind of this outcome indicator. 00:05:39.140 --> 00:05:41.720 align:middle line:90% So here are the data. 00:05:41.720 --> 00:05:43.770 align:middle line:90% And this is the trial outcome. 00:05:43.770 --> 00:05:46.420 align:middle line:84% This is what was published in this journal. 00:05:46.420 --> 00:05:48.170 align:middle line:84% And you can see that the patients assigned 00:05:48.170 --> 00:05:51.020 align:middle line:84% to the beclomethasone arm improved their FEV1 00:05:51.020 --> 00:05:54.770 align:middle line:90% about 11% to 13% on average. 00:05:54.770 --> 00:05:56.600 align:middle line:84% While the people in the montelukast group 00:05:56.600 --> 00:05:58.940 align:middle line:90% improve it 7% to 9%. 00:05:58.940 --> 00:06:01.880 align:middle line:84% Now we know from studies done in patients with asthma 00:06:01.880 --> 00:06:05.150 align:middle line:84% that if you improve your FEV1 about 10%, 00:06:05.150 --> 00:06:07.040 align:middle line:90% that you can perceive it. 00:06:07.040 --> 00:06:09.500 align:middle line:84% It's something that you say, hey, 00:06:09.500 --> 00:06:12.980 align:middle line:84% I'm better from whatever you just did. 00:06:12.980 --> 00:06:16.010 align:middle line:84% Below that, it's hard to distinguish. 00:06:16.010 --> 00:06:18.790 align:middle line:84% Above that, if you improve it twice as much, 00:06:18.790 --> 00:06:20.540 align:middle line:84% you really don't distinguish it that much, 00:06:20.540 --> 00:06:24.470 align:middle line:84% unless you happen to be engaging in really heavy exercise. 00:06:24.470 --> 00:06:26.132 align:middle line:84% So it's almost a dichotomous kind 00:06:26.132 --> 00:06:27.590 align:middle line:84% of variable, people who got better, 00:06:27.590 --> 00:06:28.875 align:middle line:90% people who didn't get better. 00:06:28.875 --> 00:06:30.500 align:middle line:84% You'd look at these data and you'd say, 00:06:30.500 --> 00:06:32.542 align:middle line:84% gee, it looks like beclomethasone is almost twice 00:06:32.542 --> 00:06:35.430 align:middle line:90% as good as montelukast. 00:06:35.430 --> 00:06:40.730 align:middle line:84% So here are the data now shown in histogram format 00:06:40.730 --> 00:06:42.680 align:middle line:84% where on the vertical axis, we have 00:06:42.680 --> 00:06:44.210 align:middle line:90% the percentage of patients. 00:06:44.210 --> 00:06:47.870 align:middle line:84% And on the horizontal axis, we have changes in FEV1. 00:06:47.870 --> 00:06:51.110 align:middle line:84% So the patients that will be over on this end 00:06:51.110 --> 00:06:53.240 align:middle line:90% had a big response. 00:06:53.240 --> 00:06:55.820 align:middle line:84% Patients over here actually got worse with treatment. 00:06:55.820 --> 00:06:58.290 align:middle line:84% And the null is right in the middle. 00:06:58.290 --> 00:07:03.950 align:middle line:84% And this 10% to 20% bar, anything over this direction, 00:07:03.950 --> 00:07:05.750 align:middle line:90% is an improvement. 00:07:05.750 --> 00:07:08.643 align:middle line:84% I don't know how to get that to go away. 00:07:08.643 --> 00:07:11.060 align:middle line:84% Usually these little boxes go away if you leave them alone 00:07:11.060 --> 00:07:11.570 align:middle line:90% long enough. 00:07:11.570 --> 00:07:16.280 align:middle line:90% 00:07:16.280 --> 00:07:17.630 align:middle line:90% Let's see if I do it again. 00:07:17.630 --> 00:07:21.370 align:middle line:90% No, cancel. 00:07:21.370 --> 00:07:22.710 align:middle line:90% So here we are with montelukast. 00:07:22.710 --> 00:07:25.050 align:middle line:84% And you see that there were about, altogether, 00:07:25.050 --> 00:07:28.500 align:middle line:84% 5% of the population who really did great. 00:07:28.500 --> 00:07:30.870 align:middle line:84% There were about 7% of the population 00:07:30.870 --> 00:07:33.270 align:middle line:90% that actually got a lot worse. 00:07:33.270 --> 00:07:36.030 align:middle line:84% And that it turns out that it's about 42% 00:07:36.030 --> 00:07:38.610 align:middle line:84% of the population that improved a significant amount. 00:07:38.610 --> 00:07:41.160 align:middle line:90% 00:07:41.160 --> 00:07:43.980 align:middle line:84% Now compare this to beclomethasone, 00:07:43.980 --> 00:07:46.330 align:middle line:84% which, remember, on average, did twice as well. 00:07:46.330 --> 00:07:47.760 align:middle line:90% Well, here's the difference. 00:07:47.760 --> 00:07:51.900 align:middle line:84% There were twice as many people that did spectacularly well. 00:07:51.900 --> 00:07:53.910 align:middle line:90% But on the average not a lot. 00:07:53.910 --> 00:07:59.070 align:middle line:84% It works out to be 10% versus 5% in some. 00:07:59.070 --> 00:08:02.280 align:middle line:84% And rather than having 7% of people who did poorly, 00:08:02.280 --> 00:08:04.300 align:middle line:90% it was just 4%. 00:08:04.300 --> 00:08:06.563 align:middle line:84% So it turns out that if you trim these 00:08:06.563 --> 00:08:08.730 align:middle line:84% by getting-- like they do sometimes in the Olympics. 00:08:08.730 --> 00:08:12.160 align:middle line:84% They get rid of your best and your worst scores. 00:08:12.160 --> 00:08:14.280 align:middle line:84% If you trim this in that end, the groups 00:08:14.280 --> 00:08:16.870 align:middle line:90% are almost identical. 00:08:16.870 --> 00:08:19.710 align:middle line:84% And so the difference in the means 00:08:19.710 --> 00:08:23.340 align:middle line:84% reflects poorly on actually what the difference 00:08:23.340 --> 00:08:25.318 align:middle line:90% would be in the population. 00:08:25.318 --> 00:08:27.360 align:middle line:84% Because if you're a patient sitting in my waiting 00:08:27.360 --> 00:08:30.581 align:middle line:84% room and your FEV1 one's improved by 60% 00:08:30.581 --> 00:08:32.039 align:middle line:84% and you're sitting next to somebody 00:08:32.039 --> 00:08:35.909 align:middle line:84% whose FEV1's improved 0%, the average improvement between you 00:08:35.909 --> 00:08:36.690 align:middle line:90% is 30%. 00:08:36.690 --> 00:08:38.909 align:middle line:84% But both of you didn't get better, 00:08:38.909 --> 00:08:41.539 align:middle line:84% even though the inference would be that you are [INAUDIBLE] 00:08:41.539 --> 00:08:42.039 align:middle line:90% better. 00:08:42.039 --> 00:08:45.750 align:middle line:84% So here we see a big population variability in response 00:08:45.750 --> 00:08:48.690 align:middle line:84% to montelukast and inhibitor of the actions of leukotrienes 00:08:48.690 --> 00:08:51.930 align:middle line:84% and leukotriene receptor in beclomethasone. 00:08:51.930 --> 00:08:54.540 align:middle line:84% So I'm going to explore the basis of these two 00:08:54.540 --> 00:09:00.000 align:middle line:84% and, depending on time, one more treatment in asthma. 00:09:00.000 --> 00:09:02.820 align:middle line:84% And there's big population variability. 00:09:02.820 --> 00:09:05.220 align:middle line:84% And the second thing that you need to know-- 00:09:05.220 --> 00:09:09.060 align:middle line:84% before I get to this, which I didn't bring you the data 00:09:09.060 --> 00:09:11.901 align:middle line:90% to show-- 00:09:11.901 --> 00:09:15.510 align:middle line:84% if I study you once, and you're over here, 00:09:15.510 --> 00:09:18.270 align:middle line:84% and I'm looking at a genetic event, that is something 00:09:18.270 --> 00:09:21.360 align:middle line:84% that you inherited, whenever I study you, 00:09:21.360 --> 00:09:25.650 align:middle line:84% you ought to get it better with this medicine. 00:09:25.650 --> 00:09:27.240 align:middle line:84% Because if it's a genetic event, it 00:09:27.240 --> 00:09:29.220 align:middle line:84% ought to be something that is imprinted on you 00:09:29.220 --> 00:09:30.870 align:middle line:90% and stay that way. 00:09:30.870 --> 00:09:34.875 align:middle line:84% While if it's variable from time to time, 00:09:34.875 --> 00:09:36.250 align:middle line:84% then it's not going to be useful. 00:09:36.250 --> 00:09:39.630 align:middle line:84% So you need to know how repeatable 00:09:39.630 --> 00:09:43.170 align:middle line:90% the treatment response is. 00:09:43.170 --> 00:09:46.680 align:middle line:84% If the basis of the variance is genetic, 00:09:46.680 --> 00:09:49.500 align:middle line:84% it ought to be highly repeatable. 00:09:49.500 --> 00:09:54.240 align:middle line:84% While if it's due to random noise like the reason 00:09:54.240 --> 00:09:56.490 align:middle line:84% you improved so much here was that 00:09:56.490 --> 00:09:58.230 align:middle line:84% during the two-week running period, 00:09:58.230 --> 00:10:00.750 align:middle line:84% you'd spent time with your mother-in-law who has six cats. 00:10:00.750 --> 00:10:02.970 align:middle line:90% And you're allergic to cats. 00:10:02.970 --> 00:10:06.720 align:middle line:84% So your lung function was low before you started. 00:10:06.720 --> 00:10:09.480 align:middle line:84% You then got better during the treatment period 00:10:09.480 --> 00:10:11.580 align:middle line:84% because you got away from the cats. 00:10:11.580 --> 00:10:13.350 align:middle line:84% If I then restudied you, you wouldn't 00:10:13.350 --> 00:10:15.660 align:middle line:90% get this kind of response. 00:10:15.660 --> 00:10:18.240 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:10:18.240 --> 00:10:21.840 align:middle line:84% JEFFREY MARK DRAZEN: Well, so actually I 00:10:21.840 --> 00:10:23.430 align:middle line:84% can show you those data, depending 00:10:23.430 --> 00:10:24.525 align:middle line:90% on how much time we have. 00:10:24.525 --> 00:10:25.775 align:middle line:90% I'll show it to you right now. 00:10:25.775 --> 00:10:27.317 align:middle line:84% It's out of the context of this talk. 00:10:27.317 --> 00:10:28.830 align:middle line:90% Just pardon me one second. 00:10:28.830 --> 00:10:29.970 align:middle line:90% I'll see if I can find-- 00:10:29.970 --> 00:10:34.515 align:middle line:84% I have these data actually in another talk. 00:10:34.515 --> 00:10:39.538 align:middle line:90% 00:10:39.538 --> 00:10:40.830 align:middle line:90% No, no, it won't take too long. 00:10:40.830 --> 00:10:42.705 align:middle line:84% Hopefully, it won't take you long to find it. 00:10:42.705 --> 00:10:44.250 align:middle line:90% could Let's see. 00:10:44.250 --> 00:10:53.070 align:middle line:90% 00:10:53.070 --> 00:10:54.690 align:middle line:90% No, it's not that one. 00:10:54.690 --> 00:10:58.830 align:middle line:90% 00:10:58.830 --> 00:11:01.950 align:middle line:90% I'll give it one more try here. 00:11:01.950 --> 00:11:03.390 align:middle line:90% Just look in a different place. 00:11:03.390 --> 00:11:18.020 align:middle line:90% 00:11:18.020 --> 00:11:19.180 align:middle line:90% I think it's here. 00:11:19.180 --> 00:11:28.960 align:middle line:90% 00:11:28.960 --> 00:11:30.310 align:middle line:90% No. 00:11:30.310 --> 00:11:32.255 align:middle line:90% One more try. 00:11:32.255 --> 00:11:33.130 align:middle line:90% Thought it was there. 00:11:33.130 --> 00:11:45.225 align:middle line:90% 00:11:45.225 --> 00:11:47.100 align:middle line:84% The nice thing about having a big hard drive, 00:11:47.100 --> 00:11:48.767 align:middle line:84% you can carry all sorts of stuff around. 00:11:48.767 --> 00:11:53.250 align:middle line:90% 00:11:53.250 --> 00:11:54.300 align:middle line:90% See if it's in this one. 00:11:54.300 --> 00:11:58.560 align:middle line:90% 00:11:58.560 --> 00:11:59.850 align:middle line:90% Yeah, here they are. 00:11:59.850 --> 00:12:05.430 align:middle line:90% OK, so here data from a trial. 00:12:05.430 --> 00:12:06.150 align:middle line:90% And I was lucky. 00:12:06.150 --> 00:12:07.800 align:middle line:84% Because this is a different drug. 00:12:07.800 --> 00:12:11.130 align:middle line:84% But this acts like a 5-lipoxygenase pathway. 00:12:11.130 --> 00:12:15.470 align:middle line:84% And the drug company that I worked with gave me the data. 00:12:15.470 --> 00:12:17.135 align:middle line:90% They sent it to me on a CD. 00:12:17.135 --> 00:12:18.510 align:middle line:84% And a lot of drug companies won't 00:12:18.510 --> 00:12:20.260 align:middle line:84% do that because they're afraid that you're 00:12:20.260 --> 00:12:23.130 align:middle line:84% going to do something that will hurt them with respect 00:12:23.130 --> 00:12:24.900 align:middle line:90% to their market position. 00:12:24.900 --> 00:12:26.310 align:middle line:90% They tend to be really paranoid. 00:12:26.310 --> 00:12:32.760 align:middle line:84% So to get on this graph, you had to improve your FEV1 by 12%. 00:12:32.760 --> 00:12:36.000 align:middle line:84% The consensus definition about what you can perceive 00:12:36.000 --> 00:12:37.590 align:middle line:90% varies between 10% and 12%. 00:12:37.590 --> 00:12:40.680 align:middle line:84% So when I did this graph, it was four or five years ago. 00:12:40.680 --> 00:12:42.480 align:middle line:90% I drew the 12% line. 00:12:42.480 --> 00:12:45.780 align:middle line:84% There were 240 people in the trial, 97 of which 00:12:45.780 --> 00:12:49.890 align:middle line:84% improved their FEV1 by 12% on the eighth day of therapy. 00:12:49.890 --> 00:12:53.430 align:middle line:84% They were then studied on day 36, 64, and 92. 00:12:53.430 --> 00:12:56.250 align:middle line:84% And this graph-- when I graphed the data on my computer, 00:12:56.250 --> 00:12:58.860 align:middle line:84% it had all these lines lying on top of each other. 00:12:58.860 --> 00:13:00.390 align:middle line:90% And they didn't get any thicker. 00:13:00.390 --> 00:13:03.790 align:middle line:84% So I can strain the data through this point here. 00:13:03.790 --> 00:13:05.070 align:middle line:90% So this isn't a data point. 00:13:05.070 --> 00:13:06.930 align:middle line:84% But this is simply to allow you to see 00:13:06.930 --> 00:13:09.570 align:middle line:84% how many people are in this group that are in that group. 00:13:09.570 --> 00:13:11.742 align:middle line:84% It's a graphical representation thing. 00:13:11.742 --> 00:13:13.950 align:middle line:84% I mean, if I had clinical data that looked like that, 00:13:13.950 --> 00:13:17.440 align:middle line:90% I wouldn't believe it. 00:13:17.440 --> 00:13:18.630 align:middle line:90% So anyhow. 00:13:18.630 --> 00:13:22.170 align:middle line:84% What you can see is that of this group of people who improved, 00:13:22.170 --> 00:13:24.840 align:middle line:84% some actually got worse and stayed worse. 00:13:24.840 --> 00:13:26.860 align:middle line:84% A lot of them got better and stayed better. 00:13:26.860 --> 00:13:29.070 align:middle line:84% And there were some that kind of moved around. 00:13:29.070 --> 00:13:31.540 align:middle line:84% If you follow the dots carefully, 00:13:31.540 --> 00:13:34.140 align:middle line:84% you'll find that three fourths of the people 00:13:34.140 --> 00:13:37.830 align:middle line:84% are above this 12% line on two of the three 00:13:37.830 --> 00:13:39.660 align:middle line:90% subsequent occasions. 00:13:39.660 --> 00:13:42.240 align:middle line:84% So that's active treatment with zileuton, 00:13:42.240 --> 00:13:43.350 align:middle line:90% which is the active drug. 00:13:43.350 --> 00:13:45.780 align:middle line:84% Here are the same entry criteria, same trial 00:13:45.780 --> 00:13:48.540 align:middle line:90% treated with placebo. 00:13:48.540 --> 00:13:50.040 align:middle line:84% And you can see that, as I told you, 00:13:50.040 --> 00:13:51.780 align:middle line:90% there are placebo responders. 00:13:51.780 --> 00:13:54.180 align:middle line:84% These people got better and stayed better. 00:13:54.180 --> 00:13:56.580 align:middle line:84% In this trial, however, of the people 00:13:56.580 --> 00:14:00.480 align:middle line:84% who were above on day eight, only about 40% 00:14:00.480 --> 00:14:02.503 align:middle line:90% are persistently above the line. 00:14:02.503 --> 00:14:04.920 align:middle line:84% And, in fact, you can see there are a bunch of people here 00:14:04.920 --> 00:14:05.740 align:middle line:90% who flunked out. 00:14:05.740 --> 00:14:07.365 align:middle line:84% And then they kind of went up and down. 00:14:07.365 --> 00:14:10.560 align:middle line:90% 00:14:10.560 --> 00:14:13.360 align:middle line:84% Now if you look at the people who don't get better. 00:14:13.360 --> 00:14:15.150 align:middle line:84% These are nonresponders to treatment, 00:14:15.150 --> 00:14:17.940 align:middle line:90% which I defined as 5% or less. 00:14:17.940 --> 00:14:20.933 align:middle line:84% So I left myself a zone of middle response. 00:14:20.933 --> 00:14:23.100 align:middle line:84% So there are a bunch of people who don't get better. 00:14:23.100 --> 00:14:24.510 align:middle line:84% And, in general, if you don't get better, 00:14:24.510 --> 00:14:25.470 align:middle line:90% you stay not better. 00:14:25.470 --> 00:14:29.940 align:middle line:84% Although there were a few people who were kind of late bloomers. 00:14:29.940 --> 00:14:31.890 align:middle line:84% It's a relatively small proportion. 00:14:31.890 --> 00:14:36.700 align:middle line:84% And when you look at placebo, it looks about the same. 00:14:36.700 --> 00:14:40.470 align:middle line:84% So you can see why the nonresponse phenotypes 00:14:40.470 --> 00:14:43.350 align:middle line:84% a little better than the response phenotype in terms 00:14:43.350 --> 00:14:44.700 align:middle line:90% of the noise. 00:14:44.700 --> 00:14:47.820 align:middle line:84% Because the people who don't get better and the placebo 00:14:47.820 --> 00:14:50.310 align:middle line:84% people who don't get better look pretty much alike. 00:14:50.310 --> 00:14:54.900 align:middle line:84% Well, the active treatment guys are contaminated by these 00:14:54.900 --> 00:14:57.928 align:middle line:84% people who you would say were better when, in fact, 00:14:57.928 --> 00:14:58.470 align:middle line:90% they weren't. 00:14:58.470 --> 00:14:59.410 align:middle line:90% So there you go. 00:14:59.410 --> 00:15:01.490 align:middle line:84% That should answer the question there. 00:15:01.490 --> 00:15:03.990 align:middle line:90% So I can close this one. 00:15:03.990 --> 00:15:13.230 align:middle line:90% 00:15:13.230 --> 00:15:16.982 align:middle line:84% I guess I closed the one I was still working on. 00:15:16.982 --> 00:15:17.660 align:middle line:90% Here you are. 00:15:17.660 --> 00:15:20.670 align:middle line:90% 00:15:20.670 --> 00:15:24.540 align:middle line:84% And so we're talking about repeatability. 00:15:24.540 --> 00:15:26.843 align:middle line:90% 00:15:26.843 --> 00:15:28.260 align:middle line:84% Have you heard about Sewall Wright 00:15:28.260 --> 00:15:31.597 align:middle line:84% and how he is a population geneticist? 00:15:31.597 --> 00:15:33.805 align:middle line:84% He did most of his work at the University of Chicago. 00:15:33.805 --> 00:15:36.120 align:middle line:84% And he died about 15, 20 years ago. 00:15:36.120 --> 00:15:40.800 align:middle line:84% He wrote a four volume series on human population genetics 00:15:40.800 --> 00:15:42.858 align:middle line:84% before we had any markers or anything. 00:15:42.858 --> 00:15:44.400 align:middle line:84% And he kind of forecast and did a lot 00:15:44.400 --> 00:15:46.590 align:middle line:84% of the primary mathematics, which 00:15:46.590 --> 00:15:50.610 align:middle line:84% are the basis of population epidemiology. 00:15:50.610 --> 00:15:54.450 align:middle line:84% And he actually derived an equation 00:15:54.450 --> 00:15:57.720 align:middle line:90% for what he calls repeatability. 00:15:57.720 --> 00:15:59.220 align:middle line:84% Now what I would really like to know 00:15:59.220 --> 00:16:01.740 align:middle line:84% is the heritability of the asthma treatment response. 00:16:01.740 --> 00:16:03.150 align:middle line:84% But to do that, I'd have to study 00:16:03.150 --> 00:16:05.130 align:middle line:90% multigenerational families. 00:16:05.130 --> 00:16:07.000 align:middle line:84% And there are three problems with that. 00:16:07.000 --> 00:16:10.350 align:middle line:84% One is that asthma treatment is changing. 00:16:10.350 --> 00:16:15.270 align:middle line:84% Second is that we know that an asthmatic at age 10, at age 25, 00:16:15.270 --> 00:16:18.930 align:middle line:84% and at age 55 responds differently to treatment. 00:16:18.930 --> 00:16:22.740 align:middle line:84% Their disease tends to get ingrained and unresponsive 00:16:22.740 --> 00:16:23.860 align:middle line:90% to treatment. 00:16:23.860 --> 00:16:25.870 align:middle line:84% And so you'd have to be able to study, 00:16:25.870 --> 00:16:29.490 align:middle line:84% let's say, all the asthmatics within a family at age 20. 00:16:29.490 --> 00:16:34.620 align:middle line:84% But the asthma treatment that you gave the patient's father 00:16:34.620 --> 00:16:36.408 align:middle line:90% is no longer available. 00:16:36.408 --> 00:16:37.450 align:middle line:90% There's something better. 00:16:37.450 --> 00:16:39.825 align:middle line:84% And so you aren't going to be able to get that treatment. 00:16:39.825 --> 00:16:42.060 align:middle line:84% And, of course, the patient's father is now 50. 00:16:42.060 --> 00:16:47.370 align:middle line:84% And you can't treat him because the asthma treatment 00:16:47.370 --> 00:16:48.390 align:middle line:90% response varies. 00:16:48.390 --> 00:16:50.770 align:middle line:84% So it's going to be almost impossible, I think, 00:16:50.770 --> 00:16:53.940 align:middle line:84% to get the data you need to determine the heritability 00:16:53.940 --> 00:16:55.710 align:middle line:90% of asthma treatment. 00:16:55.710 --> 00:16:57.890 align:middle line:90% So we look at the repeatability. 00:16:57.890 --> 00:17:01.370 align:middle line:84% And the calculated repeatability for beclomethasone 00:17:01.370 --> 00:17:03.080 align:middle line:84% and montelukast from clinical trials. 00:17:03.080 --> 00:17:06.920 align:middle line:84% And we've done on the order of 80%, which 00:17:06.920 --> 00:17:10.940 align:middle line:90% is good but not awesome. 00:17:10.940 --> 00:17:13.819 align:middle line:84% But repeatable enough we think that there's 00:17:13.819 --> 00:17:15.050 align:middle line:90% a pharmacogenetic signal. 00:17:15.050 --> 00:17:16.040 align:middle line:90% Yeah. 00:17:16.040 --> 00:17:17.859 align:middle line:84% AUDIENCE: I have just a more basic question 00:17:17.859 --> 00:17:20.609 align:middle line:84% about what [INAUDIBLE] heritability [INAUDIBLE].. 00:17:20.609 --> 00:17:24.405 align:middle line:90% 00:17:24.405 --> 00:17:25.780 align:middle line:84% JEFFREY MARK DRAZEN: So now we're 00:17:25.780 --> 00:17:27.069 align:middle line:84% not talking about the heritability 00:17:27.069 --> 00:17:28.630 align:middle line:84% of the asthmatic treatment response. 00:17:28.630 --> 00:17:32.860 align:middle line:84% We're talking about the heritability of asthma. 00:17:32.860 --> 00:17:35.920 align:middle line:84% And the asthmatic phenotype is got about-- 00:17:35.920 --> 00:17:40.360 align:middle line:84% in twin studies varies but is on the order of 65% or so, 00:17:40.360 --> 00:17:43.000 align:middle line:84% comparing monozygotic and dizygotic twins. 00:17:43.000 --> 00:17:46.839 align:middle line:84% AUDIENCE: Is there any reason to connect those two [INAUDIBLE] 00:17:46.839 --> 00:17:47.415 align:middle line:90% treatment? 00:17:47.415 --> 00:17:49.540 align:middle line:84% JEFFREY MARK DRAZEN: Well, one of the real problems 00:17:49.540 --> 00:17:53.520 align:middle line:84% with asthma is that it's difficult-- 00:17:53.520 --> 00:17:56.440 align:middle line:84% it's like an impressionistic picture. 00:17:56.440 --> 00:17:58.090 align:middle line:84% If I'm looking at it from 30 feet, 00:17:58.090 --> 00:18:00.010 align:middle line:84% it's easy to see what it's about. 00:18:00.010 --> 00:18:03.400 align:middle line:84% As I get really close up, I lose the picture. 00:18:03.400 --> 00:18:06.280 align:middle line:84% And it's very difficult to define because there's 00:18:06.280 --> 00:18:09.280 align:middle line:90% no biochemical test I can use. 00:18:09.280 --> 00:18:14.150 align:middle line:84% Quite often people look at the theorem immunoglobulin E 00:18:14.150 --> 00:18:14.650 align:middle line:90% levels. 00:18:14.650 --> 00:18:17.590 align:middle line:84% Because it's related to allergy but not asthma. 00:18:17.590 --> 00:18:20.570 align:middle line:84% But it's a very quantitative phenotype. 00:18:20.570 --> 00:18:23.530 align:middle line:84% And so the definition of asthma is airway obstruction, 00:18:23.530 --> 00:18:26.810 align:middle line:84% which varies spontaneously as a result of treatment. 00:18:26.810 --> 00:18:29.200 align:middle line:84% The problem is-- you take two people that 00:18:29.200 --> 00:18:32.410 align:middle line:84% have the same genetic background, but one of them 00:18:32.410 --> 00:18:35.560 align:middle line:84% is exposed to a lot more allergen than another. 00:18:35.560 --> 00:18:39.190 align:middle line:84% Even if you have monozygotic twins and one of them 00:18:39.190 --> 00:18:41.655 align:middle line:84% plays basketball, and changes his clothes, 00:18:41.655 --> 00:18:43.780 align:middle line:84% and wears gym clothes that have been sitting around 00:18:43.780 --> 00:18:48.190 align:middle line:84% in a ratty locker, and gets that exposure every day 00:18:48.190 --> 00:18:50.800 align:middle line:84% while the other one's a swimmer and doesn't 00:18:50.800 --> 00:18:54.700 align:middle line:84% have the same kind of exposure, one of them will develop asthma 00:18:54.700 --> 00:18:55.610 align:middle line:90% and one may not. 00:18:55.610 --> 00:18:58.510 align:middle line:84% And so you'll say it's not inheritable when, in fact, it 00:18:58.510 --> 00:19:00.400 align:middle line:84% requires both a genetic background 00:19:00.400 --> 00:19:04.150 align:middle line:84% and an environmental exposure to manifest the phenotype. 00:19:04.150 --> 00:19:06.760 align:middle line:84% The classic examples are there a lot 00:19:06.760 --> 00:19:12.790 align:middle line:84% of people from Hong Kong who have latent ragweed allergy. 00:19:12.790 --> 00:19:15.670 align:middle line:90% And they're OK in Hong Kong. 00:19:15.670 --> 00:19:18.100 align:middle line:84% There's no ragweed or very little in Hong Kong-- move 00:19:18.100 --> 00:19:19.975 align:middle line:84% to the States where there's a lot of ragweed. 00:19:19.975 --> 00:19:22.225 align:middle line:84% And they get really asthmatic in the fall. 00:19:22.225 --> 00:19:24.850 align:middle line:84% Well, they didn't have a problem back home because there wasn't 00:19:24.850 --> 00:19:29.050 align:middle line:84% the allergen. Come here and they have allergen. 00:19:29.050 --> 00:19:30.090 align:middle line:90% And they get exposed. 00:19:30.090 --> 00:19:33.670 align:middle line:84% So it's a combination of environment and genetics 00:19:33.670 --> 00:19:36.203 align:middle line:84% that appears to bring the disease on. 00:19:36.203 --> 00:19:37.870 align:middle line:84% The research I'm going to tell you about 00:19:37.870 --> 00:19:40.390 align:middle line:84% represents the work of a whole pile of people. 00:19:40.390 --> 00:19:43.622 align:middle line:84% This is most for the pile, shown here. 00:19:43.622 --> 00:19:46.080 align:middle line:84% There are people at the Channing lab, which is [? free-- ?] 00:19:46.080 --> 00:19:48.670 align:middle line:84% have heard from any of these folks, [INAUDIBLE],, 00:19:48.670 --> 00:19:50.092 align:middle line:90% I don't know. 00:19:50.092 --> 00:19:51.550 align:middle line:84% [? Zach ?] works with some of them. 00:19:51.550 --> 00:19:53.827 align:middle line:84% The Pharmacogenomics Research Network, 00:19:53.827 --> 00:19:55.660 align:middle line:84% the Asthma Clinical Research Network-- these 00:19:55.660 --> 00:20:02.312 align:middle line:84% are both NIH sponsored consortia that I've been part of. 00:20:02.312 --> 00:20:03.520 align:middle line:90% And the Whitehead Institute-- 00:20:03.520 --> 00:20:07.150 align:middle line:84% I'm sure you've all heard of those folks over at MIT. 00:20:07.150 --> 00:20:10.730 align:middle line:84% So I'm going to talk about these three asthma treatments. 00:20:10.730 --> 00:20:13.870 align:middle line:84% These are three of the major asthma treatments 00:20:13.870 --> 00:20:17.440 align:middle line:84% and what we know about their pharmacogenomics. 00:20:17.440 --> 00:20:20.150 align:middle line:90% And we'll see how things go. 00:20:20.150 --> 00:20:23.410 align:middle line:84% So with this first one, we used a candidate gene strategy. 00:20:23.410 --> 00:20:25.490 align:middle line:84% You're familiar with candidate gene strategies? 00:20:25.490 --> 00:20:29.470 align:middle line:84% So this is work I started about a decade ago. 00:20:29.470 --> 00:20:34.450 align:middle line:84% I had a post-doc from Canada who was interested in the problem. 00:20:34.450 --> 00:20:37.300 align:middle line:84% And I said, what we want to do is to look 00:20:37.300 --> 00:20:40.880 align:middle line:90% at the 5-lipoxygenase pathway. 00:20:40.880 --> 00:20:42.760 align:middle line:90% That's the leukotriene pathway. 00:20:42.760 --> 00:20:47.680 align:middle line:84% Because we knew the biochemistry of drugs. 00:20:47.680 --> 00:20:49.930 align:middle line:84% And I showed you the data, the red and green lines 00:20:49.930 --> 00:20:51.700 align:middle line:84% were from this drugs zileuton that 00:20:51.700 --> 00:20:54.380 align:middle line:84% inhibits the action of the enzyme 5-lipoxygenase. 00:20:54.380 --> 00:20:57.340 align:middle line:84% In fact, I've been involved in the development of these four 00:20:57.340 --> 00:20:59.140 align:middle line:90% drugs worldwide. 00:20:59.140 --> 00:21:01.720 align:middle line:84% And so we knew the structure of this enzyme. 00:21:01.720 --> 00:21:04.150 align:middle line:90% We knew its genomic sequence. 00:21:04.150 --> 00:21:08.680 align:middle line:84% And our question was, when you have patients with asthma who 00:21:08.680 --> 00:21:11.350 align:middle line:84% show up in your office, they all have 00:21:11.350 --> 00:21:13.360 align:middle line:90% phenotypically similar asthma. 00:21:13.360 --> 00:21:15.370 align:middle line:84% But we think that some of them may have asthma 00:21:15.370 --> 00:21:18.520 align:middle line:84% because they have an excess of 5-lipoxygenase products 00:21:18.520 --> 00:21:20.710 align:middle line:84% and stimulation at this receptor. 00:21:20.710 --> 00:21:23.230 align:middle line:84% But a clinically indistinguishable phenotype 00:21:23.230 --> 00:21:27.100 align:middle line:84% can be someone that has asthma due to an excess of substance 00:21:27.100 --> 00:21:31.310 align:middle line:84% P, a neuropeptide having nothing to do with this. 00:21:31.310 --> 00:21:34.060 align:middle line:84% Or you can have somebody that has excessive of histamine, 00:21:34.060 --> 00:21:38.590 align:middle line:84% or someone that has an excess of neurokinin A, or endothelin. 00:21:38.590 --> 00:21:40.630 align:middle line:84% So there are probably half a dozen 00:21:40.630 --> 00:21:43.600 align:middle line:84% endogenous bronchoconstrictors, which 00:21:43.600 --> 00:21:46.270 align:middle line:84% have totally different biochemical pathways that 00:21:46.270 --> 00:21:50.540 align:middle line:84% can lead to the same clinical phenotype. 00:21:50.540 --> 00:21:52.930 align:middle line:84% And so you recognize that as asthma. 00:21:52.930 --> 00:21:55.150 align:middle line:84% And they probably represent people 00:21:55.150 --> 00:21:57.550 align:middle line:84% that have a different genetically 00:21:57.550 --> 00:22:01.460 align:middle line:84% programmed mast cells in neural responses. 00:22:01.460 --> 00:22:05.770 align:middle line:84% So are argument went that if you could 00:22:05.770 --> 00:22:09.400 align:middle line:84% look at the variability of the treatment response here related 00:22:09.400 --> 00:22:11.800 align:middle line:84% to the enzymes in this pathway, you'd 00:22:11.800 --> 00:22:14.620 align:middle line:84% be able to begin to pick out people whose asthma was 00:22:14.620 --> 00:22:17.830 align:middle line:84% associated with leukotrienes and specifically whether there 00:22:17.830 --> 00:22:19.820 align:middle line:90% would be variability here. 00:22:19.820 --> 00:22:23.140 align:middle line:84% So what we started with is we had 00:22:23.140 --> 00:22:26.440 align:middle line:90% the human 5-lipoxygenase gene. 00:22:26.440 --> 00:22:28.840 align:middle line:84% We knew it's intron and exonic structure. 00:22:28.840 --> 00:22:33.263 align:middle line:84% When I started the work, this was hot, new information. 00:22:33.263 --> 00:22:35.680 align:middle line:84% And this guy from Canada worked on the problem for a month 00:22:35.680 --> 00:22:36.640 align:middle line:90% and quit. 00:22:36.640 --> 00:22:37.870 align:middle line:90% He said it was too hard. 00:22:37.870 --> 00:22:39.100 align:middle line:90% And I was really blessed. 00:22:39.100 --> 00:22:42.550 align:middle line:84% I had a post-doctoral fellow from Korea-- 00:22:42.550 --> 00:22:44.150 align:middle line:84% actually he was a visiting scientist. 00:22:44.150 --> 00:22:45.858 align:middle line:84% And he's probably the most patient person 00:22:45.858 --> 00:22:48.030 align:middle line:90% in the entire world, I think. 00:22:48.030 --> 00:22:52.740 align:middle line:84% And he plowed his way through this gene, exon by exon, 00:22:52.740 --> 00:22:58.440 align:middle line:84% doing old-fashioned sequencing, SSCP. 00:22:58.440 --> 00:23:00.390 align:middle line:84% And this took him almost three years. 00:23:00.390 --> 00:23:02.670 align:middle line:84% This kind of work now takes three weeks or maybe 00:23:02.670 --> 00:23:05.430 align:middle line:84% three days, depending on what kind of genotyping outfit 00:23:05.430 --> 00:23:06.750 align:middle line:90% you have. 00:23:06.750 --> 00:23:09.270 align:middle line:84% And what he found, which was quite discouraging, 00:23:09.270 --> 00:23:10.650 align:middle line:90% was that there were no common-- 00:23:10.650 --> 00:23:13.830 align:middle line:84% that is alleles greater than the frequency 0.15-- 00:23:13.830 --> 00:23:16.320 align:middle line:84% DNA sequence variants leading to a modified protein 00:23:16.320 --> 00:23:19.320 align:middle line:84% sequence in the entire 5-lipoxygenase gene. 00:23:19.320 --> 00:23:22.110 align:middle line:84% You got a question or are you just stretching? 00:23:22.110 --> 00:23:24.360 align:middle line:84% Wouldn't want to have a pulmonary embolism over there. 00:23:24.360 --> 00:23:28.590 align:middle line:84% So there were, however, variations 00:23:28.590 --> 00:23:30.150 align:middle line:84% in the transcription factor binding 00:23:30.150 --> 00:23:32.250 align:middle line:90% region, which is just upstream. 00:23:32.250 --> 00:23:34.300 align:middle line:90% It's the 5-LO gene promoter. 00:23:34.300 --> 00:23:37.150 align:middle line:84% Now what was known about the gene promoter at that time, 00:23:37.150 --> 00:23:39.630 align:middle line:84% which has actually hung in there to be true, 00:23:39.630 --> 00:23:42.810 align:middle line:84% is it just the head of the translation start site, 00:23:42.810 --> 00:23:47.520 align:middle line:84% there are a series of SP1 and EGR1 binding motifs. 00:23:47.520 --> 00:23:52.590 align:middle line:84% In fact, there are five of them in tandem. 00:23:52.590 --> 00:23:55.560 align:middle line:84% Then if you search the GenBank, this 00:23:55.560 --> 00:23:58.420 align:middle line:84% is the only gene where that occurs. 00:23:58.420 --> 00:24:01.380 align:middle line:84% So the sequence GGGCGG is repeated over and over 00:24:01.380 --> 00:24:02.910 align:middle line:90% and over again. 00:24:02.910 --> 00:24:06.137 align:middle line:84% And that's just right in the core promoter. 00:24:06.137 --> 00:24:08.220 align:middle line:84% And that's where the sequence variants were found. 00:24:08.220 --> 00:24:12.090 align:middle line:84% In fact, the sequence in the GenBank 00:24:12.090 --> 00:24:14.820 align:middle line:90% has five of these in a row. 00:24:14.820 --> 00:24:21.330 align:middle line:84% And we've identified individuals with 3, 4, and 6, and 7, less 00:24:21.330 --> 00:24:24.240 align:middle line:84% commonly, we've, in fact, identified people with 2. 00:24:24.240 --> 00:24:27.030 align:middle line:84% And so this is a variable nucleotide tandem repeat. 00:24:27.030 --> 00:24:29.010 align:middle line:84% People could use it for genotyping. 00:24:29.010 --> 00:24:31.830 align:middle line:90% But we showed that it read true. 00:24:31.830 --> 00:24:33.990 align:middle line:84% So it's not one that it tends to expand. 00:24:33.990 --> 00:24:37.860 align:middle line:84% It's not like it is in Huntington's disease. 00:24:37.860 --> 00:24:41.922 align:middle line:84% If your mother and father have different alleles 00:24:41.922 --> 00:24:43.380 align:middle line:84% and are homozygous for them, you're 00:24:43.380 --> 00:24:45.150 align:middle line:90% going to be a heterozygous. 00:24:45.150 --> 00:24:47.730 align:middle line:84% So we did enough family studies to know 00:24:47.730 --> 00:24:50.640 align:middle line:90% that this was a stable VNTR. 00:24:50.640 --> 00:24:56.640 align:middle line:84% And then what we did was we took human cells-- 00:24:56.640 --> 00:24:58.260 align:middle line:90% these are HeLa cells-- 00:24:58.260 --> 00:25:02.130 align:middle line:84% and transected them with the wild type promoter 00:25:02.130 --> 00:25:05.370 align:middle line:84% and promoters which has five repeats, 00:25:05.370 --> 00:25:09.450 align:middle line:90% and with 3, 4, and 6 repeats. 00:25:09.450 --> 00:25:11.240 align:middle line:84% And studied them in a CAT assay, this 00:25:11.240 --> 00:25:13.720 align:middle line:84% is a chloramphenicol acetyltransferase assay. 00:25:13.720 --> 00:25:17.473 align:middle line:84% So the higher the bar, the greater the promoter activity. 00:25:17.473 --> 00:25:19.140 align:middle line:84% And you can see that all the mutants are 00:25:19.140 --> 00:25:22.500 align:middle line:84% less effective at driving gene transcription 00:25:22.500 --> 00:25:24.480 align:middle line:90% than the wild type. 00:25:24.480 --> 00:25:27.507 align:middle line:84% Now this is one of these moments that you never forget 00:25:27.507 --> 00:25:28.590 align:middle line:90% in clinical investigation. 00:25:28.590 --> 00:25:31.800 align:middle line:84% I was going over these data with Tucker Collins, 00:25:31.800 --> 00:25:34.338 align:middle line:84% who's a professor of pathology at Brigham now. 00:25:34.338 --> 00:25:35.880 align:middle line:84% And he looked at these data and says, 00:25:35.880 --> 00:25:37.260 align:middle line:84% you know, Drazen, you're absolutely right. 00:25:37.260 --> 00:25:38.802 align:middle line:84% But you've got to find something else 00:25:38.802 --> 00:25:43.710 align:middle line:84% to do because these variations are too small. 00:25:43.710 --> 00:25:46.380 align:middle line:84% No one wants to study a 35% or 40% 00:25:46.380 --> 00:25:52.110 align:middle line:84% decrease in CAT transferase activities, too small. 00:25:52.110 --> 00:25:54.450 align:middle line:84% But what I knew from clinical trials 00:25:54.450 --> 00:25:59.100 align:middle line:84% was that if I inhibited 5-lipoxygenase 35% or 40%, 00:25:59.100 --> 00:26:02.890 align:middle line:84% that I got a very substantial clinical benefit. 00:26:02.890 --> 00:26:05.700 align:middle line:84% So the argument that we had in our head 00:26:05.700 --> 00:26:10.410 align:middle line:84% was that if patients with variations 00:26:10.410 --> 00:26:15.210 align:middle line:84% in the 5-lipoxygenase, the ALOX5 promoter, 00:26:15.210 --> 00:26:19.410 align:middle line:84% had downregulated ALOX5 production, 00:26:19.410 --> 00:26:23.800 align:middle line:84% but they wouldn't respond to antileukotriene treatment. 00:26:23.800 --> 00:26:27.600 align:middle line:84% The idea is that patients with this funny form 00:26:27.600 --> 00:26:32.550 align:middle line:84% of the core promoter would have some other cause 00:26:32.550 --> 00:26:33.390 align:middle line:90% for their asthma. 00:26:33.390 --> 00:26:39.330 align:middle line:84% Their asthma may be due to the histamines, or to substance P, 00:26:39.330 --> 00:26:43.830 align:middle line:84% or some other variant, some other mediator, 00:26:43.830 --> 00:26:46.230 align:middle line:84% but that leukotrienes wouldn't be important. 00:26:46.230 --> 00:26:49.090 align:middle line:84% Because these are all downregulatory mutations. 00:26:49.090 --> 00:26:53.010 align:middle line:84% So I was lucky, at the time, because Abbott Pharmaceuticals 00:26:53.010 --> 00:26:56.610 align:middle line:84% had developed the 5-lipoxygenase inhibitors zileuton. 00:26:56.610 --> 00:26:59.940 align:middle line:84% And it had two big problems they were trying to overcome. 00:26:59.940 --> 00:27:05.250 align:middle line:84% One is that it has to be given to patients four times a day. 00:27:05.250 --> 00:27:08.340 align:middle line:90% And that's a pain. 00:27:08.340 --> 00:27:12.150 align:middle line:84% If you give patients the pill to take once a day, they'll do it. 00:27:12.150 --> 00:27:13.890 align:middle line:90% Twice a day is harder. 00:27:13.890 --> 00:27:16.020 align:middle line:84% Three times a day requires a saint. 00:27:16.020 --> 00:27:17.850 align:middle line:84% And four times a day requires God. 00:27:17.850 --> 00:27:22.470 align:middle line:90% It's almost impossible to do it. 00:27:22.470 --> 00:27:24.990 align:middle line:84% Now I did actually have patients that had pretty bad asthma 00:27:24.990 --> 00:27:26.290 align:middle line:90% that took it four times a day. 00:27:26.290 --> 00:27:29.730 align:middle line:84% So Abbott developed the compound you could take twice a day. 00:27:29.730 --> 00:27:34.080 align:middle line:84% The second problem for them was that about 3% 00:27:34.080 --> 00:27:37.350 align:middle line:84% of people that took this drug developed abnormalities 00:27:37.350 --> 00:27:39.540 align:middle line:90% in liver function tests. 00:27:39.540 --> 00:27:44.250 align:middle line:84% And although nobody had a liver that checked out on this drug, 00:27:44.250 --> 00:27:47.115 align:middle line:84% it required monitoring of the whole population 00:27:47.115 --> 00:27:49.315 align:middle line:84% to find the 3% or 4% of people who had 00:27:49.315 --> 00:27:50.840 align:middle line:90% an adverse hepatic response. 00:27:50.840 --> 00:27:54.130 align:middle line:84% It's a very common problem in drug development. 00:27:54.130 --> 00:27:55.720 align:middle line:84% And so they came up with a daughter 00:27:55.720 --> 00:27:58.610 align:middle line:84% of zileuton, which wasn't supposed to have this problem. 00:27:58.610 --> 00:28:00.700 align:middle line:90% And it was called ABT 761. 00:28:00.700 --> 00:28:02.560 align:middle line:90% It acted at the same location. 00:28:02.560 --> 00:28:04.190 align:middle line:90% And that was the plan. 00:28:04.190 --> 00:28:06.670 align:middle line:84% So it was the clinical trial that we designed. 00:28:06.670 --> 00:28:09.020 align:middle line:90% I actually helped design it. 00:28:09.020 --> 00:28:12.520 align:middle line:84% This was not a high-tech clinical trial. 00:28:12.520 --> 00:28:14.560 align:middle line:84% We enrolled patients on no treatment. 00:28:14.560 --> 00:28:17.890 align:middle line:84% And we assigned them to-- these were asthma patients-- 00:28:17.890 --> 00:28:21.100 align:middle line:90% to ABT-761 or a placebo. 00:28:21.100 --> 00:28:23.620 align:middle line:84% And since they were patients that we found on no treatment, 00:28:23.620 --> 00:28:27.130 align:middle line:84% continuing them on no treatment seemed like it was no problem. 00:28:27.130 --> 00:28:28.960 align:middle line:84% Now this was seven or eight years ago. 00:28:28.960 --> 00:28:30.520 align:middle line:90% Asthma treatments progressed. 00:28:30.520 --> 00:28:32.950 align:middle line:84% It would be harder to do this trial now. 00:28:32.950 --> 00:28:35.413 align:middle line:84% Because there are a lot of people who didn't get treatment 00:28:35.413 --> 00:28:36.080 align:middle line:90% [? who would. ?] 00:28:36.080 --> 00:28:38.770 align:middle line:84% So what happened was we genotyped everybody 00:28:38.770 --> 00:28:40.180 align:middle line:90% at the 5-lipoxygenase locus. 00:28:40.180 --> 00:28:41.990 align:middle line:90% And as we expected-- 00:28:41.990 --> 00:28:44.950 align:middle line:84% oh, I forgot to tell you one more thing which 00:28:44.950 --> 00:28:48.730 align:middle line:84% was that when the trial was set up, Abbott, to save money, 00:28:48.730 --> 00:28:51.307 align:middle line:84% said, look, when a third of the patients complete-- 00:28:51.307 --> 00:28:53.140 align:middle line:84% you don't enroll everybody in the first day. 00:28:53.140 --> 00:28:55.473 align:middle line:84% Usually, it takes a couple of years to enroll-- 00:28:55.473 --> 00:28:56.890 align:middle line:84% but when the third of the patients 00:28:56.890 --> 00:28:58.390 align:middle line:84% complete their 12 weeks, we're going 00:28:58.390 --> 00:29:01.150 align:middle line:84% to compare the incidence of liver function tests 00:29:01.150 --> 00:29:03.550 align:middle line:84% in the placebo and ABT-761 group. 00:29:03.550 --> 00:29:07.400 align:middle line:84% And if this drug isn't improving the rate of adverse effects, 00:29:07.400 --> 00:29:08.650 align:middle line:90% we're going to stop the trial. 00:29:08.650 --> 00:29:11.260 align:middle line:84% Because it's not worth another $40 million 00:29:11.260 --> 00:29:13.900 align:middle line:84% to find out something we don't want to know, 00:29:13.900 --> 00:29:15.710 align:middle line:84% which was that this drug wasn't in advance. 00:29:15.710 --> 00:29:17.710 align:middle line:84% Because they were hoping it would be in advance. 00:29:17.710 --> 00:29:20.410 align:middle line:84% And I thought that was a very reasonable approach to life. 00:29:20.410 --> 00:29:21.920 align:middle line:90% So we genotyped everybody. 00:29:21.920 --> 00:29:23.830 align:middle line:84% And as you might expect, the wild type allele 00:29:23.830 --> 00:29:27.580 align:middle line:84% occurred most commonly, made up about 81% of the alleles. 00:29:27.580 --> 00:29:32.140 align:middle line:84% The deletion alleles three and four made up 19%. 00:29:32.140 --> 00:29:35.830 align:middle line:84% The addition allele six was so rare-- 00:29:35.830 --> 00:29:38.950 align:middle line:84% with one patient with it-- that it really 00:29:38.950 --> 00:29:40.990 align:middle line:90% didn't add up to much. 00:29:40.990 --> 00:29:45.190 align:middle line:84% So when we stratified the results of FEV1 00:29:45.190 --> 00:29:50.170 align:middle line:84% by genotype, the patients that received active treatment 00:29:50.170 --> 00:29:53.290 align:middle line:84% and where the wild type genotype, 55, 00:29:53.290 --> 00:29:57.070 align:middle line:90% improved their FEV1 about 17%. 00:29:57.070 --> 00:30:03.250 align:middle line:84% Patients on placebo with the 55 genotype improved FEV1 5%. 00:30:03.250 --> 00:30:04.720 align:middle line:84% Now it turned out that there were 00:30:04.720 --> 00:30:09.680 align:middle line:84% 14 patients that had active treatment that had no 5 allele. 00:30:09.680 --> 00:30:11.260 align:middle line:90% So we called them active XX. 00:30:11.260 --> 00:30:14.770 align:middle line:84% And they actually had their FEV1 get worse. 00:30:14.770 --> 00:30:17.860 align:middle line:84% And since we didn't stratify the enrollment by genotype, 00:30:17.860 --> 00:30:20.290 align:middle line:84% it turned out there was only one patient that 00:30:20.290 --> 00:30:22.855 align:middle line:90% had XX genotype on placebo. 00:30:22.855 --> 00:30:23.980 align:middle line:90% And that person got better. 00:30:23.980 --> 00:30:25.923 align:middle line:90% But that's one patient. 00:30:25.923 --> 00:30:27.590 align:middle line:84% It's hard to make any sense out of that. 00:30:27.590 --> 00:30:31.870 align:middle line:84% So the p-value for this change in FEV1 00:30:31.870 --> 00:30:34.090 align:middle line:90% was 10 to the minus fifth. 00:30:34.090 --> 00:30:37.160 align:middle line:84% This is unlikely to have happened by chance. 00:30:37.160 --> 00:30:42.130 align:middle line:84% But when we went into the trial based on a preliminary analysis 00:30:42.130 --> 00:30:44.800 align:middle line:84% of the data set that we never published because we were 00:30:44.800 --> 00:30:49.180 align:middle line:84% concerned about bias in ascertaining the patients, 00:30:49.180 --> 00:30:51.850 align:middle line:84% we had thought that heterozygosity at this locus 00:30:51.850 --> 00:30:54.018 align:middle line:84% would contribute to the phenotype. 00:30:54.018 --> 00:30:55.810 align:middle line:84% And it turned out that that's not the case. 00:30:55.810 --> 00:30:57.430 align:middle line:90% You had to be homozygous. 00:30:57.430 --> 00:31:00.740 align:middle line:84% So that the allele frequency here is 0.19 squared. 00:31:00.740 --> 00:31:01.780 align:middle line:90% So it's very small. 00:31:01.780 --> 00:31:04.460 align:middle line:84% It's about 3% or 4% of the population. 00:31:04.460 --> 00:31:07.720 align:middle line:84% So from a pharmacogenetic perspective, it's interesting, 00:31:07.720 --> 00:31:11.180 align:middle line:84% but it's not economically interesting. 00:31:11.180 --> 00:31:13.580 align:middle line:84% On the other hand, if it had been 20% of the population, 00:31:13.580 --> 00:31:14.872 align:middle line:90% it would have been interesting. 00:31:14.872 --> 00:31:17.290 align:middle line:84% Because if you assume that you could do a genotype for $50 00:31:17.290 --> 00:31:18.143 align:middle line:90% bucks. 00:31:18.143 --> 00:31:19.810 align:middle line:84% It's easier to treat the patient and see 00:31:19.810 --> 00:31:23.470 align:middle line:84% if they get better than genotype 100 people to find people 00:31:23.470 --> 00:31:26.260 align:middle line:84% three people who aren't going to get better. 00:31:26.260 --> 00:31:27.850 align:middle line:84% Since there are also probably other 00:31:27.850 --> 00:31:30.058 align:middle line:84% reasons that you don't get better with this treatment 00:31:30.058 --> 00:31:31.520 align:middle line:90% other than this genotype. 00:31:31.520 --> 00:31:34.405 align:middle line:84% So it turned out to be pharmacoeconomically 00:31:34.405 --> 00:31:35.280 align:middle line:90% not very interesting. 00:31:35.280 --> 00:31:39.910 align:middle line:84% Now it turned out that some work we did at the same time 00:31:39.910 --> 00:31:42.460 align:middle line:90% was that we got eosinophils. 00:31:42.460 --> 00:31:44.800 align:middle line:84% And we look for the expression of mRNA 00:31:44.800 --> 00:31:46.900 align:middle line:90% by PCR of the ALOX5 gene. 00:31:46.900 --> 00:31:49.240 align:middle line:84% This is our control with [INAUDIBLE].. 00:31:49.240 --> 00:31:54.190 align:middle line:84% Then we had five patients here with a 55 genotype. 00:31:54.190 --> 00:31:56.920 align:middle line:84% In most of them, but not all of them, 00:31:56.920 --> 00:31:59.120 align:middle line:84% have relatively high levels of the five 00:31:59.120 --> 00:32:02.890 align:middle line:84% well picked up ALOX5 picked up by PCR. 00:32:02.890 --> 00:32:07.060 align:middle line:84% But none of the three patients that we did that had no 5 00:32:07.060 --> 00:32:10.450 align:middle line:90% allele had this availability. 00:32:10.450 --> 00:32:14.350 align:middle line:84% And then we looked at the amount of LPC-4, 00:32:14.350 --> 00:32:18.370 align:middle line:84% that is the leukotrienes produced by these NFLs. 00:32:18.370 --> 00:32:21.730 align:middle line:84% The deletional variance-- there were four patients 00:32:21.730 --> 00:32:24.123 align:middle line:90% for that compared to five here. 00:32:24.123 --> 00:32:25.540 align:middle line:84% There was a significant difference 00:32:25.540 --> 00:32:29.350 align:middle line:84% in terms of lesser amounts of leukotrienes produced. 00:32:29.350 --> 00:32:35.260 align:middle line:84% So that suggested that there was variation in this pathway. 00:32:35.260 --> 00:32:37.480 align:middle line:84% We've since looked at the cytosolic phospholipase 00:32:37.480 --> 00:32:40.150 align:middle line:84% A2, the LPC-4 synthase, the epoxide hydrolase, 00:32:40.150 --> 00:32:42.280 align:middle line:84% all the other genes in this pathway. 00:32:42.280 --> 00:32:44.280 align:middle line:84% And we haven't found any other variance 00:32:44.280 --> 00:32:45.530 align:middle line:90% related to treatment response. 00:32:45.530 --> 00:32:48.100 align:middle line:90% 00:32:48.100 --> 00:32:49.990 align:middle line:84% But we did find-- and this is actually 00:32:49.990 --> 00:32:51.730 align:middle line:90% interesting confirmatory data. 00:32:51.730 --> 00:32:54.880 align:middle line:84% And this is not our work but the work of others. 00:32:54.880 --> 00:32:59.560 align:middle line:84% What they reasoned was that if the variance that 00:32:59.560 --> 00:33:02.080 align:middle line:84% changed the amount of look leukotriene 00:33:02.080 --> 00:33:05.590 align:middle line:84% produced, that you should get the same pharmacogenetic effect 00:33:05.590 --> 00:33:08.200 align:middle line:84% if patients are treated with one of these drugs, which 00:33:08.200 --> 00:33:11.650 align:middle line:84% rather than inhibiting the enzyme are receptor inhibitors. 00:33:11.650 --> 00:33:14.850 align:middle line:90% 00:33:14.850 --> 00:33:22.570 align:middle line:84% In fact, I'll tell you that in the trials done 00:33:22.570 --> 00:33:25.630 align:middle line:84% with the receptor inhibitors by others, 00:33:25.630 --> 00:33:28.000 align:middle line:84% patients with the mutant form of the genotype 00:33:28.000 --> 00:33:31.000 align:middle line:84% don't respond to the receptor inhibitor, 00:33:31.000 --> 00:33:32.200 align:middle line:90% so confirming our data. 00:33:32.200 --> 00:33:33.940 align:middle line:90% It's always heart-warming. 00:33:33.940 --> 00:33:36.670 align:middle line:84% So it suggests that variations in this pathway 00:33:36.670 --> 00:33:39.890 align:middle line:84% can contribute to a small fraction of the variance. 00:33:39.890 --> 00:33:42.220 align:middle line:84% So I think it's interesting pharmacogenetics. 00:33:42.220 --> 00:33:45.250 align:middle line:84% But it never made it pharmacoeconomically 00:33:45.250 --> 00:33:49.450 align:middle line:84% because it's a too small proportion of the population. 00:33:49.450 --> 00:33:52.480 align:middle line:84% We really have to, I think, affect 15%-- 00:33:52.480 --> 00:33:56.270 align:middle line:84% unless you're talking about a serious toxic event. 00:33:56.270 --> 00:34:02.470 align:middle line:84% Now, for example, with the gene thiopurine methyltransferase, 00:34:02.470 --> 00:34:08.020 align:middle line:84% which is a gene involved in the metabolism of 6-mercaptopurine, 00:34:08.020 --> 00:34:11.199 align:middle line:84% which is a drug used to treat leukemia. 00:34:11.199 --> 00:34:13.810 align:middle line:84% Among members of the population, especially people 00:34:13.810 --> 00:34:19.570 align:middle line:84% of Scandinavian heritage, about 2% or 3% of those people 00:34:19.570 --> 00:34:21.310 align:middle line:84% harbor an allele where they don't 00:34:21.310 --> 00:34:26.449 align:middle line:84% metabolize 6-mercaptopurine at the rate they should. 00:34:26.449 --> 00:34:28.870 align:middle line:84% And so if you give them the standard dose, they get toxic. 00:34:28.870 --> 00:34:33.639 align:middle line:84% And you can actually kill them with the drug. 00:34:33.639 --> 00:34:39.010 align:middle line:84% So when you're treating leukemia now with that drug, 00:34:39.010 --> 00:34:41.020 align:middle line:84% it's standard to genotype people to make 00:34:41.020 --> 00:34:44.530 align:middle line:84% sure they don't have one of these slow-metabolizing 00:34:44.530 --> 00:34:45.580 align:middle line:90% alleles. 00:34:45.580 --> 00:34:48.190 align:middle line:84% And they're now you're looking for a life threatening 00:34:48.190 --> 00:34:51.429 align:middle line:84% toxicity, which occurs at low frequency. 00:34:51.429 --> 00:34:56.260 align:middle line:84% But because the consequences of missing it 00:34:56.260 --> 00:35:01.000 align:middle line:84% are so dramatic and so irreversible, 00:35:01.000 --> 00:35:03.580 align:middle line:84% it makes sense to screen 1,000 people to find one. 00:35:03.580 --> 00:35:05.770 align:middle line:84% In fact, if you screen someone in Asia, 00:35:05.770 --> 00:35:09.550 align:middle line:84% you have to screen almost 5,000 people to find one. 00:35:09.550 --> 00:35:11.290 align:middle line:84% But yet, in treating leukemia in Asia, 00:35:11.290 --> 00:35:14.050 align:middle line:84% they're still doing it, just because it's 00:35:14.050 --> 00:35:16.720 align:middle line:84% something that can be prevented if you're going to a higher 00:35:16.720 --> 00:35:18.460 align:middle line:90% class treatment center. 00:35:18.460 --> 00:35:20.590 align:middle line:84% Well, if you were screening at the Mayo Clinic 00:35:20.590 --> 00:35:23.020 align:middle line:84% where there's a lot of Scandinavian blood, 00:35:23.020 --> 00:35:24.830 align:middle line:90% really make sense to do it. 00:35:24.830 --> 00:35:28.480 align:middle line:84% But for something like this, where the failure 00:35:28.480 --> 00:35:31.915 align:middle line:84% to respond to treatment is a nuisance as 00:35:31.915 --> 00:35:33.415 align:middle line:84% opposed to a life-threatening event, 00:35:33.415 --> 00:35:36.440 align:middle line:84% it didn't make pharmacoeconomic sense. 00:35:36.440 --> 00:35:40.840 align:middle line:84% So my second example is about inhaled steroids. 00:35:40.840 --> 00:35:44.110 align:middle line:84% And steroids are different from antileukotrienes 00:35:44.110 --> 00:35:46.330 align:middle line:84% because we knew how antileukotrienes 00:35:46.330 --> 00:35:47.620 align:middle line:90% work in asthma. 00:35:47.620 --> 00:35:50.270 align:middle line:84% I mean, they were developed based on understanding 00:35:50.270 --> 00:35:51.730 align:middle line:90% of the biology of asthma. 00:35:51.730 --> 00:35:55.210 align:middle line:90% 00:35:55.210 --> 00:35:58.510 align:middle line:84% But inhaled steroids we knew worked based on observation. 00:35:58.510 --> 00:36:02.710 align:middle line:84% But steroids have thousands of potential mechanisms of action. 00:36:02.710 --> 00:36:06.470 align:middle line:84% We had no idea which one was accurate. 00:36:06.470 --> 00:36:10.150 align:middle line:84% So some of this study is useful because if I 00:36:10.150 --> 00:36:14.350 align:middle line:84% can find a gene that controls steroid response, 00:36:14.350 --> 00:36:16.120 align:middle line:84% maybe that's telling me a gene that's 00:36:16.120 --> 00:36:18.730 align:middle line:84% important in the biology of asthma. 00:36:18.730 --> 00:36:20.920 align:middle line:84% And then, I could develop a treatment 00:36:20.920 --> 00:36:24.070 align:middle line:84% that inhibited just that gene and wouldn't 00:36:24.070 --> 00:36:26.500 align:middle line:84% have a lot of the side effects of steroids. 00:36:26.500 --> 00:36:30.130 align:middle line:84% Because one of the big problems with treating asthma 00:36:30.130 --> 00:36:33.640 align:middle line:84% with steroids is that if you're a kid, it makes you shorter. 00:36:33.640 --> 00:36:36.610 align:middle line:84% If you're an adult, it makes your bones brittle. 00:36:36.610 --> 00:36:39.640 align:middle line:84% It causes acne to break out, causes skin to thin. 00:36:39.640 --> 00:36:42.940 align:middle line:84% It's got a lot of side effects that aren't terribly good. 00:36:42.940 --> 00:36:47.350 align:middle line:84% And to get rid of those side effects 00:36:47.350 --> 00:36:49.390 align:middle line:84% would be good until you can get at it. 00:36:49.390 --> 00:36:51.760 align:middle line:84% At the same time, with inhaled steroids, 00:36:51.760 --> 00:36:53.380 align:middle line:84% just like with antileukotrienes, you 00:36:53.380 --> 00:36:57.190 align:middle line:84% saw that there are about 40% of people who are paying for drugs 00:36:57.190 --> 00:36:59.290 align:middle line:90% who aren't getting any benefit. 00:36:59.290 --> 00:37:01.790 align:middle line:84% But the drug companies think that's perfectly fine. 00:37:01.790 --> 00:37:02.800 align:middle line:90% I don't. 00:37:02.800 --> 00:37:06.310 align:middle line:84% Because they're getting the toxic effects, we know. 00:37:06.310 --> 00:37:08.320 align:middle line:84% But they're not getting the therapeutic benefit. 00:37:08.320 --> 00:37:11.800 align:middle line:84% And it's costing them $50 bucks a month, sometimes more. 00:37:11.800 --> 00:37:13.450 align:middle line:84% The new inhaled steroids are actually 00:37:13.450 --> 00:37:15.010 align:middle line:90% $60 to $75 bucks a month. 00:37:15.010 --> 00:37:16.900 align:middle line:90% So it's an expensive habit. 00:37:16.900 --> 00:37:19.750 align:middle line:84% It's a lot like your coffee habit. 00:37:19.750 --> 00:37:23.440 align:middle line:84% If you buy a couple of cups of coffee at the Starbucks 00:37:23.440 --> 00:37:25.210 align:middle line:84% or even at the place down here, you're 00:37:25.210 --> 00:37:27.793 align:middle line:84% spending two 2 and 1/2 bucks a day on coffee. 00:37:27.793 --> 00:37:28.960 align:middle line:90% Works out $70 bucks a month. 00:37:28.960 --> 00:37:31.470 align:middle line:84% That's what an asthma attack costs. 00:37:31.470 --> 00:37:34.660 align:middle line:84% So here we are showing two more populations. 00:37:34.660 --> 00:37:38.410 align:middle line:84% I showed you the purple population 00:37:38.410 --> 00:37:41.410 align:middle line:84% with the beclomethasone There are two 00:37:41.410 --> 00:37:43.780 align:middle line:90% different inhaled steroids-- 00:37:43.780 --> 00:37:46.980 align:middle line:84% different from the ones I showed you-- showing a variation 00:37:46.980 --> 00:37:49.852 align:middle line:84% treatment response quite similar to the one 00:37:49.852 --> 00:37:51.560 align:middle line:84% that I showed you from that asthma trial. 00:37:51.560 --> 00:37:54.550 align:middle line:84% Now, these two populations are ones that we studied. 00:37:54.550 --> 00:37:56.880 align:middle line:84% This is actually a drug company study 00:37:56.880 --> 00:38:01.540 align:middle line:84% done by Forest Labs of one of their inhaled steroids. 00:38:01.540 --> 00:38:03.780 align:middle line:84% And this was one sponsored by the NIH called 00:38:03.780 --> 00:38:07.740 align:middle line:84% the Childhood Asthma Management Program or CAMP. 00:38:07.740 --> 00:38:12.180 align:middle line:84% And even though ones kids and ones adults, totally different 00:38:12.180 --> 00:38:15.120 align:middle line:84% study designs, you see this same variation 00:38:15.120 --> 00:38:17.880 align:middle line:84% in the asthma treatment response to inhaled steroids 00:38:17.880 --> 00:38:19.710 align:middle line:84% in three populations, allowing one 00:38:19.710 --> 00:38:22.870 align:middle line:84% to believe it's highly likely to be true. 00:38:22.870 --> 00:38:25.890 align:middle line:84% So what we did here is we have a different candidate gene 00:38:25.890 --> 00:38:26.400 align:middle line:90% strategy. 00:38:26.400 --> 00:38:30.370 align:middle line:84% And the 5-lipoxygenase-- we knew the genes in the pathway. 00:38:30.370 --> 00:38:34.350 align:middle line:84% And we could identify them based on how the treatment worked. 00:38:34.350 --> 00:38:37.680 align:middle line:84% Here, we were guessing how the treatment worked. 00:38:37.680 --> 00:38:40.528 align:middle line:84% So we got a bunch of people around in a room who 00:38:40.528 --> 00:38:41.820 align:middle line:90% we thought knew about steroids. 00:38:41.820 --> 00:38:45.540 align:middle line:84% And we compiled the list of possible genes in the pathway. 00:38:45.540 --> 00:38:49.560 align:middle line:84% We then looked for DNA sequence variance in that pathway. 00:38:49.560 --> 00:38:51.960 align:middle line:84% We then look to see if there's a statistical relationship 00:38:51.960 --> 00:38:54.450 align:middle line:84% between clinical response and the presence of the sequence 00:38:54.450 --> 00:38:55.380 align:middle line:90% variant. 00:38:55.380 --> 00:38:57.540 align:middle line:84% And then we determined their functional relevance. 00:38:57.540 --> 00:39:01.440 align:middle line:84% And so I don't need to define SNPs, right? 00:39:01.440 --> 00:39:03.510 align:middle line:84% You all know about SNPs and haplotypes, right? 00:39:03.510 --> 00:39:05.320 align:middle line:90% Talked all about that. 00:39:05.320 --> 00:39:06.435 align:middle line:90% So here's our strategy. 00:39:06.435 --> 00:39:09.160 align:middle line:90% 00:39:09.160 --> 00:39:14.190 align:middle line:84% We have 32 control and 16 asthmatic cell lines, 00:39:14.190 --> 00:39:17.700 align:middle line:84% which we had an infinite amount of DNA from. 00:39:17.700 --> 00:39:23.370 align:middle line:84% And we identified variants from sequencing and from a database. 00:39:23.370 --> 00:39:27.450 align:middle line:84% We then selected SNPs based on the allele frequency. 00:39:27.450 --> 00:39:31.050 align:middle line:84% We designed genotyping assays and did primary genotyping 00:39:31.050 --> 00:39:34.260 align:middle line:84% in the adult study-- that was the drug company study-- 00:39:34.260 --> 00:39:36.135 align:middle line:90% of cases and controls. 00:39:36.135 --> 00:39:46.860 align:middle line:90% 00:39:46.860 --> 00:39:48.650 align:middle line:84% What we were using is cases and controls 00:39:48.650 --> 00:39:50.690 align:middle line:84% was comparing people at this end of the graph 00:39:50.690 --> 00:39:52.890 align:middle line:84% to people at that end of the graph. 00:39:52.890 --> 00:39:56.315 align:middle line:84% So we call these controls and these cases or vice versa. 00:39:56.315 --> 00:40:00.730 align:middle line:90% 00:40:00.730 --> 00:40:04.410 align:middle line:84% And then we would identify haplotype tag SNPs that 00:40:04.410 --> 00:40:06.810 align:middle line:90% had greater than 5% prevalence. 00:40:06.810 --> 00:40:08.610 align:middle line:84% And that gave us first [INAUDIBLE] 00:40:08.610 --> 00:40:12.280 align:middle line:84% [? statistical ?] associations in the adult population. 00:40:12.280 --> 00:40:15.150 align:middle line:84% We then said if this is true, we should be able to replicate it 00:40:15.150 --> 00:40:17.550 align:middle line:90% in the kids. 00:40:17.550 --> 00:40:19.950 align:middle line:84% And then we said if it's true still, 00:40:19.950 --> 00:40:23.350 align:middle line:84% should be able to replicate it in a third population. 00:40:23.350 --> 00:40:27.160 align:middle line:84% And then we did some fancy statistics on it. 00:40:27.160 --> 00:40:31.620 align:middle line:84% So here are the three populations we studied. 00:40:31.620 --> 00:40:33.780 align:middle line:84% This adult study is a drug company study, 00:40:33.780 --> 00:40:36.360 align:middle line:84% eight weeks in duration, people whose lung 00:40:36.360 --> 00:40:39.695 align:middle line:84% function as a percent [INAUDIBLE] is 70% to start. 00:40:39.695 --> 00:40:40.320 align:middle line:90% They're adults. 00:40:40.320 --> 00:40:42.360 align:middle line:90% They're age 40. 00:40:42.360 --> 00:40:44.100 align:middle line:90% And they're interesting. 00:40:44.100 --> 00:40:48.270 align:middle line:84% Their steroid response is 7% on average. 00:40:48.270 --> 00:40:52.200 align:middle line:84% In the CAMP trial, which is our first replicate, they're kids. 00:40:52.200 --> 00:40:55.110 align:middle line:90% So their average age is 9. 00:40:55.110 --> 00:40:58.710 align:middle line:84% The trial duration was four years not 00:40:58.710 --> 00:41:00.330 align:middle line:90% four weeks or four months. 00:41:00.330 --> 00:41:04.140 align:middle line:84% And their average improvement FEV1 is 7%. 00:41:04.140 --> 00:41:06.250 align:middle line:84% And then our second replication trial-- 00:41:06.250 --> 00:41:08.910 align:middle line:84% was government sponsored, six weeks. 00:41:08.910 --> 00:41:12.130 align:middle line:84% Again, average improvements about 7%. 00:41:12.130 --> 00:41:14.310 align:middle line:84% So this is about what you get in these populations. 00:41:14.310 --> 00:41:17.880 align:middle line:84% This kind of population, you get about 7% improvement FEV1. 00:41:17.880 --> 00:41:21.960 align:middle line:84% This is the list of genes that we interrogated. 00:41:21.960 --> 00:41:25.110 align:middle line:84% And this was simply a list of genes made up 00:41:25.110 --> 00:41:27.900 align:middle line:84% by understanding the pathway, thinking 00:41:27.900 --> 00:41:30.340 align:middle line:90% how it could impinge on asthma. 00:41:30.340 --> 00:41:33.400 align:middle line:90% So it's entirely theoretical. 00:41:33.400 --> 00:41:39.600 align:middle line:84% And so here our primary outcome using the single SNP analysis-- 00:41:39.600 --> 00:41:42.570 align:middle line:90% these are haplotype tag SNPs. 00:41:42.570 --> 00:41:47.190 align:middle line:84% So in the adult study, RS-242941, 00:41:47.190 --> 00:41:50.820 align:middle line:84% so this is just the name of one of these haplotype tag SNPs-- 00:41:50.820 --> 00:41:53.670 align:middle line:84% using the eight-week FEV1 percent change 00:41:53.670 --> 00:42:01.950 align:middle line:84% as a continuous variable, gave us a p-value of 0.025. 00:42:01.950 --> 00:42:04.710 align:middle line:84% This is adjusted for covariates but not 00:42:04.710 --> 00:42:09.120 align:middle line:84% for the number of [? looks. ?] We then did it in a CAMP 00:42:09.120 --> 00:42:13.920 align:middle line:84% population, now many fewer looks, again 00:42:13.920 --> 00:42:16.350 align:middle line:84% eight-week percent change, continuous variable, 00:42:16.350 --> 00:42:19.270 align:middle line:90% p-value of 0.006. 00:42:19.270 --> 00:42:20.550 align:middle line:90% So you're on a roll. 00:42:20.550 --> 00:42:24.720 align:middle line:84% And we looked at that same genotype in the ACRN population 00:42:24.720 --> 00:42:29.880 align:middle line:90% and didn't find a relationship. 00:42:29.880 --> 00:42:34.020 align:middle line:84% That would have been hitting gold. 00:42:34.020 --> 00:42:36.660 align:middle line:84% So we got silver or some other load. 00:42:36.660 --> 00:42:38.970 align:middle line:84% I'm not quite sure what you want to call it. 00:42:38.970 --> 00:42:45.150 align:middle line:84% But we found that these are three other haplotype tag SNPs. 00:42:45.150 --> 00:42:54.330 align:middle line:84% That this one, RS-1876828, which is in the same gene, CRHR1, 00:42:54.330 --> 00:42:59.850 align:middle line:84% cortical troponin releasing hormone receptor 1, 00:42:59.850 --> 00:43:02.350 align:middle line:84% was positive in both populations. 00:43:02.350 --> 00:43:04.350 align:middle line:84% So we got either second or third prize, 00:43:04.350 --> 00:43:06.150 align:middle line:90% depending on your perspective. 00:43:06.150 --> 00:43:09.270 align:middle line:84% The ideal would have been to have the same haplotype tags 00:43:09.270 --> 00:43:12.420 align:middle line:84% SNP positive in all three populations. 00:43:12.420 --> 00:43:14.820 align:middle line:84% What we found was different haplotype tags 00:43:14.820 --> 00:43:20.910 align:middle line:84% SNPs in the same gene, suggesting that it's 00:43:20.910 --> 00:43:26.460 align:middle line:84% the gene that's the problem but not localizing the site, 00:43:26.460 --> 00:43:30.820 align:middle line:84% not actually getting to what we think is the active problem. 00:43:30.820 --> 00:43:36.150 align:middle line:84% So the half-full way of looking at these data 00:43:36.150 --> 00:43:39.090 align:middle line:84% are that, gee, although I don't know the mechanism, 00:43:39.090 --> 00:43:44.070 align:middle line:84% if I genotype people at these loci, 00:43:44.070 --> 00:43:48.460 align:middle line:84% can I predict their response in lung function? 00:43:48.460 --> 00:43:54.720 align:middle line:90% So this is the RS-242941. 00:43:54.720 --> 00:43:58.860 align:middle line:84% That was the top one in the kids CAMP and the adults. 00:43:58.860 --> 00:44:03.240 align:middle line:84% And what we've not done is taken this haplotype tag SNP 00:44:03.240 --> 00:44:06.850 align:middle line:84% and looked at people who were homozygous for it, 00:44:06.850 --> 00:44:10.680 align:middle line:84% and heterozygous, and didn't have any copies of it. 00:44:10.680 --> 00:44:13.500 align:middle line:84% And you can see that if you own this phenotype, 00:44:13.500 --> 00:44:17.850 align:middle line:84% that your improvement in FEV1 is almost twice as great as if you 00:44:17.850 --> 00:44:21.210 align:middle line:90% own this genotype on average. 00:44:21.210 --> 00:44:26.430 align:middle line:84% When you look at a different haplotype. 00:44:26.430 --> 00:44:28.860 align:middle line:90% This is GAT haplotype pair. 00:44:28.860 --> 00:44:32.720 align:middle line:84% And this is now in the adult and the CAMP study. 00:44:32.720 --> 00:44:34.800 align:middle line:84% Even it's a little more informative here. 00:44:34.800 --> 00:44:37.590 align:middle line:84% I can show you the difference in the size of these bars. 00:44:37.590 --> 00:44:41.520 align:middle line:84% Here, this is about a 10% to 15%. 00:44:41.520 --> 00:44:44.840 align:middle line:84% If you look here, this is now 15% to 20%. 00:44:44.840 --> 00:44:48.000 align:middle line:84% This is a little bigger response in people with this haplotype 00:44:48.000 --> 00:44:49.960 align:middle line:90% compared to people without it. 00:44:49.960 --> 00:44:54.030 align:middle line:84% So we think that it has some predictive value. 00:44:54.030 --> 00:44:56.570 align:middle line:84% It did not turn out to be statistically significant, 00:44:56.570 --> 00:44:59.590 align:middle line:84% as I said again, in the ACRN population, 00:44:59.590 --> 00:45:00.910 align:middle line:90% smaller number of people. 00:45:00.910 --> 00:45:05.300 align:middle line:84% It was in the right sign but not statistically significant. 00:45:05.300 --> 00:45:09.070 align:middle line:84% This is a different genotype in that same gene 00:45:09.070 --> 00:45:12.640 align:middle line:84% but a different SNP, the RS-1876828, 00:45:12.640 --> 00:45:17.560 align:middle line:84% showing again that if you're homozygous AA versus GG, 00:45:17.560 --> 00:45:19.970 align:middle line:90% a big difference in response. 00:45:19.970 --> 00:45:23.600 align:middle line:84% But when we went back and did this population 00:45:23.600 --> 00:45:27.650 align:middle line:84% with that genotype, it didn't work out. 00:45:27.650 --> 00:45:30.700 align:middle line:84% The second thing we learned is that CRHR1 is probably 00:45:30.700 --> 00:45:33.490 align:middle line:84% involved in the of the asthma treatment response, 00:45:33.490 --> 00:45:35.170 align:middle line:84% suggesting that if you could target 00:45:35.170 --> 00:45:39.760 align:middle line:84% CRHR1 antagonist to the lung, where CRHR1 receptor. 00:45:39.760 --> 00:45:41.980 align:middle line:84% It's a seven-transmembrane spanning G protein 00:45:41.980 --> 00:45:43.270 align:middle line:90% coupled receptor. 00:45:43.270 --> 00:45:45.100 align:middle line:84% It's got a bunch of ligands in addition 00:45:45.100 --> 00:45:50.450 align:middle line:84% to corticotropin releasing hormone that can activate it. 00:45:50.450 --> 00:45:53.480 align:middle line:84% So some of the questions are, can you find it in the lung? 00:45:53.480 --> 00:45:55.720 align:middle line:84% So this is one of these body blots 00:45:55.720 --> 00:46:01.300 align:middle line:84% that you buy from one of those companies out west. 00:46:01.300 --> 00:46:04.330 align:middle line:84% [? Say, ?] oh, [? it's ?] RNA, if you do, 00:46:04.330 --> 00:46:08.130 align:middle line:84% what we did here is 12 cycles of PCR. 00:46:08.130 --> 00:46:09.880 align:middle line:84% And you pick it up in the positive control 00:46:09.880 --> 00:46:11.120 align:middle line:90% and in the brain. 00:46:11.120 --> 00:46:12.590 align:middle line:90% This is a hypothalamic hormone. 00:46:12.590 --> 00:46:15.048 align:middle line:84% So the fact that you find the hypothalamic hormone receptor 00:46:15.048 --> 00:46:16.570 align:middle line:90% in the brain is no surprise. 00:46:16.570 --> 00:46:19.300 align:middle line:84% If you do twice as much PCR, you not only 00:46:19.300 --> 00:46:20.590 align:middle line:90% pick it up in the brain. 00:46:20.590 --> 00:46:24.130 align:middle line:84% It shows up in the lung, in the placenta, the thymus, 00:46:24.130 --> 00:46:25.870 align:middle line:90% and the lymph nodes. 00:46:25.870 --> 00:46:29.230 align:middle line:84% It shows up in epithelial cells and CD4 positive lymphocytes, 00:46:29.230 --> 00:46:31.210 align:middle line:84% all tissues that have been implicated 00:46:31.210 --> 00:46:33.260 align:middle line:90% in the biology of asthma. 00:46:33.260 --> 00:46:35.080 align:middle line:84% So at least it's where it needs to be. 00:46:35.080 --> 00:46:37.250 align:middle line:84% Doesn't quite meet Koch's postulates yet. 00:46:37.250 --> 00:46:41.840 align:middle line:84% But we're headed in the right direction. 00:46:41.840 --> 00:46:44.650 align:middle line:84% For the third example, what I think I'll talk about 00:46:44.650 --> 00:46:46.280 align:middle line:90% is a beta-agonist receptor. 00:46:46.280 --> 00:46:50.020 align:middle line:84% So what I've done so far is talk about two association studies. 00:46:50.020 --> 00:46:52.480 align:middle line:90% Clinical trials that were done. 00:46:52.480 --> 00:46:55.090 align:middle line:84% We then get the data and the DNA and we 00:46:55.090 --> 00:46:57.340 align:middle line:84% do a bunch of genotyping and associations 00:46:57.340 --> 00:47:01.480 align:middle line:90% but it's all post hoc. 00:47:01.480 --> 00:47:03.100 align:middle line:84% So now the story of beta agonists 00:47:03.100 --> 00:47:05.500 align:middle line:90% are slightly different. 00:47:05.500 --> 00:47:08.080 align:middle line:84% Beta-agonist inhalers are the most commonly used asthma 00:47:08.080 --> 00:47:09.998 align:middle line:90% treatment in the world. 00:47:09.998 --> 00:47:12.040 align:middle line:84% I don't know how many of you read the New England 00:47:12.040 --> 00:47:12.850 align:middle line:90% Journal of Medicine. 00:47:12.850 --> 00:47:14.933 align:middle line:84% Have ever see the New England Journal of Medicine? 00:47:14.933 --> 00:47:15.850 align:middle line:90% Nice graphics. 00:47:15.850 --> 00:47:17.510 align:middle line:84% [INAUDIBLE] lollipops in this picture. 00:47:17.510 --> 00:47:21.410 align:middle line:84% She's one of the artists that draws them. 00:47:21.410 --> 00:47:25.840 align:middle line:84% And so that these inhalers are going off worldwide 00:47:25.840 --> 00:47:29.840 align:middle line:84% at the rate of about 1,000 times a second. 00:47:29.840 --> 00:47:31.510 align:middle line:84% So there's a lot of-- treat people 00:47:31.510 --> 00:47:33.710 align:middle line:90% using this kind of treatment. 00:47:33.710 --> 00:47:40.390 align:middle line:84% So in 1990, almost 15 years ago, this fire was started. 00:47:40.390 --> 00:47:42.615 align:middle line:84% A long-acting beta agonist came out. 00:47:42.615 --> 00:47:43.990 align:middle line:84% Now the problem with those little 00:47:43.990 --> 00:47:46.910 align:middle line:84% inhalers-- they last three or four hours. 00:47:46.910 --> 00:47:51.130 align:middle line:84% So this drug company designed a drug that can last and agonise 00:47:51.130 --> 00:47:54.190 align:middle line:90% at the receptor for 12 hours. 00:47:54.190 --> 00:47:57.335 align:middle line:84% And the hypothesis was that if you took it on a regular basis, 00:47:57.335 --> 00:47:58.960 align:middle line:84% your asthma would improve because you'd 00:47:58.960 --> 00:48:01.450 align:middle line:90% be chronically dilated. 00:48:01.450 --> 00:48:04.750 align:middle line:84% These drugs work by relaxing airway smooth muscle 00:48:04.750 --> 00:48:07.040 align:middle line:90% constricted in asthma. 00:48:07.040 --> 00:48:12.490 align:middle line:84% So in this trial designed by the sponsors, 00:48:12.490 --> 00:48:14.860 align:middle line:84% they were comparing this long-acting beta-agonist 00:48:14.860 --> 00:48:15.730 align:middle line:90% with placebo. 00:48:15.730 --> 00:48:17.840 align:middle line:90% And it was a crossover design. 00:48:17.840 --> 00:48:20.260 align:middle line:84% So half the patients started on formoterol, half 00:48:20.260 --> 00:48:23.560 align:middle line:84% started on placebo for 26 weeks of treatment. 00:48:23.560 --> 00:48:25.000 align:middle line:90% They got a month off. 00:48:25.000 --> 00:48:27.610 align:middle line:84% And then they were switch to the other treatment 00:48:27.610 --> 00:48:30.670 align:middle line:84% and followed for another 26 weeks. 00:48:30.670 --> 00:48:32.980 align:middle line:84% And in this trial, the primary outcome 00:48:32.980 --> 00:48:35.500 align:middle line:90% was an asthma exacerbation. 00:48:35.500 --> 00:48:37.193 align:middle line:84% Rather than measuring lung function, 00:48:37.193 --> 00:48:39.610 align:middle line:84% they said I'm going to wait for your asthma get bad enough 00:48:39.610 --> 00:48:42.010 align:middle line:90% for you to be upset by it. 00:48:42.010 --> 00:48:44.710 align:middle line:84% And I'm going to click off the asthma exacerbation 00:48:44.710 --> 00:48:48.160 align:middle line:90% box when that happens. 00:48:48.160 --> 00:48:51.100 align:middle line:84% In the hypothesis had been that during your final treatment, 00:48:51.100 --> 00:48:52.960 align:middle line:84% you'd have fewer asthma exacerbations 00:48:52.960 --> 00:48:55.660 align:middle line:84% than during your placebo treatment period. 00:48:55.660 --> 00:48:59.680 align:middle line:84% Because of the crossover design, everybody's in both arms. 00:48:59.680 --> 00:49:02.290 align:middle line:84% So when these data were published, 00:49:02.290 --> 00:49:04.840 align:middle line:90% the asthma world was surprised. 00:49:04.840 --> 00:49:08.090 align:middle line:84% Because this is the number of subjects without exacerbation 00:49:08.090 --> 00:49:09.445 align:middle line:90% with time. 00:49:09.445 --> 00:49:14.440 align:middle line:84% If treatment was 100% effective, your line would go across here. 00:49:14.440 --> 00:49:17.590 align:middle line:84% Every time a patient has an asthma exacerbation, 00:49:17.590 --> 00:49:18.880 align:middle line:90% the line kicks down. 00:49:18.880 --> 00:49:21.850 align:middle line:84% And the patient's then censored from further analysis. 00:49:21.850 --> 00:49:24.370 align:middle line:84% And you could see that there for about a week 00:49:24.370 --> 00:49:26.170 align:middle line:90% or so they were the same. 00:49:26.170 --> 00:49:29.740 align:middle line:84% But then the patients that got the regular treatment-- 00:49:29.740 --> 00:49:32.260 align:middle line:84% remember, the onset prevention was worth a pound of cure. 00:49:32.260 --> 00:49:34.480 align:middle line:84% This is supposed to make you better but actually 00:49:34.480 --> 00:49:38.110 align:middle line:84% having more asthma exacerbations at any given point in time 00:49:38.110 --> 00:49:40.330 align:middle line:84% than the people who only used beta-agonist 00:49:40.330 --> 00:49:42.380 align:middle line:90% when they needed it. 00:49:42.380 --> 00:49:45.730 align:middle line:84% And when these data came out, there 00:49:45.730 --> 00:49:50.200 align:middle line:84% was concern that actually this very commonly used asthma 00:49:50.200 --> 00:49:54.340 align:middle line:90% treatment was causing harm. 00:49:54.340 --> 00:49:57.520 align:middle line:84% And the fires were fueled by this paper 00:49:57.520 --> 00:49:59.255 align:middle line:84% that was published by my predecessors 00:49:59.255 --> 00:50:00.380 align:middle line:90% at the New England Journal. 00:50:00.380 --> 00:50:04.391 align:middle line:84% I had nothing to do with this, let you know. 00:50:04.391 --> 00:50:06.410 align:middle line:84% showing that the use of beta-agonists 00:50:06.410 --> 00:50:08.570 align:middle line:84% was associated with the risk of death or near death 00:50:08.570 --> 00:50:09.060 align:middle line:90% through asthma. 00:50:09.060 --> 00:50:10.477 align:middle line:84% And what they did in this study is 00:50:10.477 --> 00:50:13.820 align:middle line:84% they went to this huge database, which is where every time you 00:50:13.820 --> 00:50:16.730 align:middle line:84% get a prescription in the province of Saskatchewan, 00:50:16.730 --> 00:50:20.480 align:middle line:84% somebody knows who you are and what prescription you want. 00:50:20.480 --> 00:50:24.770 align:middle line:84% And then they also matched that with the asthma exacerbations, 00:50:24.770 --> 00:50:28.670 align:middle line:84% asthma death file when people got really bad asthma. 00:50:28.670 --> 00:50:30.830 align:middle line:84% Because they have hospital records. 00:50:30.830 --> 00:50:33.020 align:middle line:84% And they discovered that the patients that 00:50:33.020 --> 00:50:37.070 align:middle line:84% used a lot of beta-agonist either died from their asthma 00:50:37.070 --> 00:50:40.310 align:middle line:84% or were admitted to the hospital with very severe asthma. 00:50:40.310 --> 00:50:44.340 align:middle line:84% And they concluded it was cause and effect. 00:50:44.340 --> 00:50:46.430 align:middle line:90% And they were totally wrong. 00:50:46.430 --> 00:50:50.870 align:middle line:84% This is a classic example of confounding by severity. 00:50:50.870 --> 00:50:53.150 align:middle line:84% Or more simply put, the sicker patients 00:50:53.150 --> 00:50:55.190 align:middle line:84% were using their beta inhalers a lot. 00:50:55.190 --> 00:50:58.980 align:middle line:84% And they're the ones that are more likely to die from asthma. 00:50:58.980 --> 00:51:01.400 align:middle line:90% So it was bad epidemiology. 00:51:01.400 --> 00:51:04.400 align:middle line:84% But while we were working that out, 00:51:04.400 --> 00:51:06.530 align:middle line:84% this trial was conceived by the people 00:51:06.530 --> 00:51:10.010 align:middle line:84% at the NIH, who became concerned that a common asthma 00:51:10.010 --> 00:51:12.420 align:middle line:90% medicine was making you sick. 00:51:12.420 --> 00:51:14.390 align:middle line:84% And so we put together this trial 00:51:14.390 --> 00:51:16.820 align:middle line:90% called the beta-agonist study. 00:51:16.820 --> 00:51:19.315 align:middle line:84% All clinical trials have to have a name. 00:51:19.315 --> 00:51:20.690 align:middle line:84% Otherwise, you would have to say, 00:51:20.690 --> 00:51:23.480 align:middle line:84% remember that trial where they compared drug X to drug Y 00:51:23.480 --> 00:51:26.060 align:middle line:84% in a 24-week crossover design trial. 00:51:26.060 --> 00:51:29.630 align:middle line:90% It was published in The Lancet. 00:51:29.630 --> 00:51:31.980 align:middle line:84% remember the bags trial or something like that. 00:51:31.980 --> 00:51:34.340 align:middle line:84% So all good clinical trials have a name. 00:51:34.340 --> 00:51:36.180 align:middle line:90% This was the bags trial. 00:51:36.180 --> 00:51:37.850 align:middle line:84% And in this trial we enrolled patients 00:51:37.850 --> 00:51:40.910 align:middle line:84% with really mild asthma for a six-week run-in period 00:51:40.910 --> 00:51:43.220 align:middle line:84% when they were on their standard treatment. 00:51:43.220 --> 00:51:46.190 align:middle line:84% And then they were randomized to receive either albuterol, which 00:51:46.190 --> 00:51:48.620 align:middle line:90% is a short-acting beta inhaler. 00:51:48.620 --> 00:51:50.510 align:middle line:90% So you take it four times a day. 00:51:50.510 --> 00:51:55.280 align:middle line:84% This is the one that's used so often around the world. 00:51:55.280 --> 00:51:56.990 align:middle line:84% And they were given a coated inhaler. 00:51:56.990 --> 00:51:58.700 align:middle line:90% It was white. 00:51:58.700 --> 00:51:59.780 align:middle line:90% And it said, study drug. 00:51:59.780 --> 00:52:02.360 align:middle line:90% Take two puffs four times a day. 00:52:02.360 --> 00:52:04.860 align:middle line:84% And I've already told you about four times a day medication. 00:52:04.860 --> 00:52:08.390 align:middle line:84% So this inhaler had in it a computer chip. 00:52:08.390 --> 00:52:09.710 align:middle line:90% And the patients knew this. 00:52:09.710 --> 00:52:11.690 align:middle line:84% Every time they pressed the inhaler, 00:52:11.690 --> 00:52:14.960 align:middle line:84% it recorded the time of day and the date. 00:52:14.960 --> 00:52:17.780 align:middle line:84% And we knew that we had about 95% 00:52:17.780 --> 00:52:20.510 align:middle line:84% compliance, at least with pressing the inhaler. 00:52:20.510 --> 00:52:22.730 align:middle line:84% There wasn't a video camera in the computer chip. 00:52:22.730 --> 00:52:25.220 align:middle line:84% I don't know whether the patient actually inhaled 00:52:25.220 --> 00:52:26.720 align:middle line:90% when they pressed the inhaler. 00:52:26.720 --> 00:52:29.030 align:middle line:84% But we think that pressing the inhaler 00:52:29.030 --> 00:52:32.328 align:middle line:84% meant that they actually used it. 00:52:32.328 --> 00:52:33.620 align:middle line:90% And that was a coating inhaler. 00:52:33.620 --> 00:52:36.530 align:middle line:84% And they were also given an open-label albuterol. 00:52:36.530 --> 00:52:38.870 align:middle line:84% And said, if your asthma is still acting up, 00:52:38.870 --> 00:52:41.400 align:middle line:90% you can take this. 00:52:41.400 --> 00:52:43.580 align:middle line:84% And so all the patients were given that. 00:52:43.580 --> 00:52:46.910 align:middle line:84% The placebo group was given an identical inhaler, 00:52:46.910 --> 00:52:49.160 align:middle line:84% identical instructions, identical computer chip, 00:52:49.160 --> 00:52:51.410 align:middle line:90% and open-label albuterol. 00:52:51.410 --> 00:52:55.310 align:middle line:84% And the hypothesis was that if regular use of beta-agonist 00:52:55.310 --> 00:52:58.010 align:middle line:84% was bad for you, that the blue group would do poorly 00:52:58.010 --> 00:53:00.260 align:middle line:90% compared to the green group. 00:53:00.260 --> 00:53:03.230 align:middle line:84% Now rather than powering the study for asthma exacerbations, 00:53:03.230 --> 00:53:05.480 align:middle line:84% because in these very mild asthma patients, 00:53:05.480 --> 00:53:09.470 align:middle line:84% you were taking 1,000 patients studied for a year. 00:53:09.470 --> 00:53:11.030 align:middle line:84% We used a surrogate endpoint, which 00:53:11.030 --> 00:53:13.580 align:middle line:84% was lung function, strongly related 00:53:13.580 --> 00:53:23.340 align:middle line:84% to how patients do with asthma, taken over 16 weeks. 00:53:23.340 --> 00:53:25.430 align:middle line:84% And here the data which we published actually 00:53:25.430 --> 00:53:29.630 align:middle line:84% in the New England Journal of Medicine in '96. 00:53:29.630 --> 00:53:32.390 align:middle line:84% It showed that the morning peak, which 00:53:32.390 --> 00:53:36.260 align:middle line:84% is our primary outcome, the two groups 00:53:36.260 --> 00:53:38.180 align:middle line:84% were not statistically distinguishable. 00:53:38.180 --> 00:53:40.378 align:middle line:84% Here, it's a very expanded scale. 00:53:40.378 --> 00:53:42.920 align:middle line:84% And the difference between these two and 15 liters per minute 00:53:42.920 --> 00:53:45.570 align:middle line:90% is probably not significant. 00:53:45.570 --> 00:53:48.920 align:middle line:84% We had gone in saying that this difference, 25, was 00:53:48.920 --> 00:53:49.910 align:middle line:90% significant. 00:53:49.910 --> 00:53:52.200 align:middle line:90% And it got only halfway there. 00:53:52.200 --> 00:53:54.530 align:middle line:84% But interestingly, the blue group 00:53:54.530 --> 00:53:56.570 align:middle line:84% did a little poorer than the green group, 00:53:56.570 --> 00:53:57.945 align:middle line:84% even though the difference wasn't 00:53:57.945 --> 00:53:59.310 align:middle line:90% statistically significant. 00:53:59.310 --> 00:54:02.840 align:middle line:84% This wasn't a sign of these treatments being worse for you. 00:54:02.840 --> 00:54:04.940 align:middle line:84% But we concluded that they weren't worse for you. 00:54:04.940 --> 00:54:07.260 align:middle line:84% In fact, we looked at a bunch of outcomes. 00:54:07.260 --> 00:54:10.760 align:middle line:84% Now while we were doing this trial, while data accrual was 00:54:10.760 --> 00:54:12.710 align:middle line:90% going on-- 00:54:12.710 --> 00:54:15.200 align:middle line:84% this is the beta-2 adrenergic receptor-- 00:54:15.200 --> 00:54:18.680 align:middle line:84% a number of polymorphisms were identified in it. 00:54:18.680 --> 00:54:20.960 align:middle line:84% Now these two, which are right up here 00:54:20.960 --> 00:54:24.050 align:middle line:84% near the n-terminal part of the receptor, 00:54:24.050 --> 00:54:25.910 align:middle line:84% have very high allele frequencies. 00:54:25.910 --> 00:54:29.225 align:middle line:84% The minor allele frequencies are on the order of 0.4, 00:54:29.225 --> 00:54:31.010 align:middle line:90% and they're very common. 00:54:31.010 --> 00:54:32.630 align:middle line:84% And we knew from the work of others 00:54:32.630 --> 00:54:34.100 align:middle line:90% that these were functional. 00:54:34.100 --> 00:54:37.250 align:middle line:84% That is that an [? RH16 ?] receptor behaved differently 00:54:37.250 --> 00:54:39.680 align:middle line:90% than a [? glycine ?] receptor. 00:54:39.680 --> 00:54:42.260 align:middle line:84% We also knew that RH16 was in very 00:54:42.260 --> 00:54:46.620 align:middle line:84% strong linkage disequilibrium with [? Glyn ?] 27. 00:54:46.620 --> 00:54:50.420 align:middle line:84% So that if you were RH16, you were likely Glenn 27. 00:54:50.420 --> 00:54:53.450 align:middle line:84% So what we did was we stratified the data 00:54:53.450 --> 00:54:58.370 align:middle line:84% from that clinical trial based on these genotypes, again, 00:54:58.370 --> 00:55:00.590 align:middle line:90% a retrospective analysis. 00:55:00.590 --> 00:55:01.910 align:middle line:90% And here are the findings. 00:55:01.910 --> 00:55:07.790 align:middle line:84% If you had the Gly/Gly genotype and used beta-agonist 00:55:07.790 --> 00:55:10.610 align:middle line:84% on a regular basis, your morning peak 00:55:10.610 --> 00:55:16.070 align:middle line:84% flow after treatment oscillates around the zero line. 00:55:16.070 --> 00:55:20.330 align:middle line:84% If you have the Arg/Arg genotype and only used beta-agonist 00:55:20.330 --> 00:55:22.880 align:middle line:84% when you needed it, which works out to be about a puff 00:55:22.880 --> 00:55:26.600 align:middle line:84% every other day, nothing happened. 00:55:26.600 --> 00:55:29.900 align:middle line:84% But if you have the Arg/Arg genotype 00:55:29.900 --> 00:55:33.000 align:middle line:84% and are using two puffs four times a day, 00:55:33.000 --> 00:55:36.050 align:middle line:84% there's a fall in peak flow during and after 00:55:36.050 --> 00:55:39.680 align:middle line:84% the active treatment period, suggesting 00:55:39.680 --> 00:55:43.020 align:middle line:84% that it's not that the drugs don't work anymore. 00:55:43.020 --> 00:55:45.860 align:middle line:84% In fact, when you take them you get bronchodilation. 00:55:45.860 --> 00:55:47.570 align:middle line:90% But there's a side effect. 00:55:47.570 --> 00:55:50.240 align:middle line:90% 00:55:50.240 --> 00:55:53.360 align:middle line:84% When the drug wears off, you're worse 00:55:53.360 --> 00:55:57.272 align:middle line:84% than you were before you started taking it. 00:55:57.272 --> 00:55:58.730 align:middle line:84% And that's what this is showing us. 00:55:58.730 --> 00:56:02.280 align:middle line:84% Because this is the morning peak flow before treatment. 00:56:02.280 --> 00:56:05.090 align:middle line:84% So this says that using a drug on a regular basis 00:56:05.090 --> 00:56:07.610 align:middle line:84% makes you worse in the morning than you would 00:56:07.610 --> 00:56:09.420 align:middle line:90% be if you hadn't used the drug. 00:56:09.420 --> 00:56:11.000 align:middle line:90% But it's genotype related. 00:56:11.000 --> 00:56:13.860 align:middle line:90% 00:56:13.860 --> 00:56:17.840 align:middle line:84% So that led us to do a perspective trial. 00:56:17.840 --> 00:56:20.930 align:middle line:84% And so this is like the gold standard in this business. 00:56:20.930 --> 00:56:23.840 align:middle line:84% Rather than sifting through all data, 00:56:23.840 --> 00:56:25.460 align:middle line:84% you start off with new data that were 00:56:25.460 --> 00:56:29.120 align:middle line:84% designed by BARG trial, beta-agonist 00:56:29.120 --> 00:56:31.580 align:middle line:90% receptor by genotype. 00:56:31.580 --> 00:56:35.510 align:middle line:84% And this was two identical trials 00:56:35.510 --> 00:56:40.370 align:middle line:84% in which patients were enrolled with the Arg/Arg genotype. 00:56:40.370 --> 00:56:42.770 align:middle line:84% A matching Gly/Gly genotype patient 00:56:42.770 --> 00:56:47.270 align:middle line:84% was found, at least with respect to sex and lung function. 00:56:47.270 --> 00:56:52.250 align:middle line:84% And then both groups of patients were put in this design. 00:56:52.250 --> 00:56:55.790 align:middle line:84% And the idea was that in the Arg/Arg patients, 00:56:55.790 --> 00:57:01.700 align:middle line:84% we were expecting that placebo would be superior to albuterol. 00:57:01.700 --> 00:57:04.850 align:middle line:84% Because we think that these are the patients where albuterol 00:57:04.850 --> 00:57:08.150 align:middle line:84% given regularly has a detrimental effect. 00:57:08.150 --> 00:57:12.560 align:middle line:84% While in the Gly/Gly patients, this difference 00:57:12.560 --> 00:57:14.430 align:middle line:90% would not exist. 00:57:14.430 --> 00:57:16.760 align:middle line:84% And then if we then did a genotype 00:57:16.760 --> 00:57:19.980 align:middle line:90% by treatment interaction-- 00:57:19.980 --> 00:57:23.450 align:middle line:84% we would say that if you look at Gly/Gly on active treatment 00:57:23.450 --> 00:57:25.850 align:middle line:84% versus Arg/Arg on active treatment, 00:57:25.850 --> 00:57:28.130 align:middle line:84% that these patients would be better, 00:57:28.130 --> 00:57:29.880 align:middle line:84% these patients would be worse, and there'd 00:57:29.880 --> 00:57:32.880 align:middle line:84% be a big difference between the two groups. 00:57:32.880 --> 00:57:34.370 align:middle line:90% And that's how we designed it. 00:57:34.370 --> 00:57:36.080 align:middle line:84% Our primary outcome variable, again, 00:57:36.080 --> 00:57:39.920 align:middle line:84% was a measurement of lung function, morning peak flow, 00:57:39.920 --> 00:57:41.780 align:middle line:90% with these secondary outcomes. 00:57:41.780 --> 00:57:44.770 align:middle line:90% 00:57:44.770 --> 00:57:48.930 align:middle line:84% So we enrolled 332 patients with that genotype. 00:57:48.930 --> 00:57:51.660 align:middle line:90% 78 met the criteria. 00:57:51.660 --> 00:57:53.160 align:middle line:84% And so this is an interesting trial. 00:57:53.160 --> 00:57:56.130 align:middle line:84% Because as far as I know, it's the first trial 00:57:56.130 --> 00:57:59.610 align:middle line:84% where people were enrolled by genotype 00:57:59.610 --> 00:58:01.770 align:middle line:90% in a non-malignant condition. 00:58:01.770 --> 00:58:04.320 align:middle line:84% They came in and they met clinical and physiological 00:58:04.320 --> 00:58:05.100 align:middle line:90% criteria. 00:58:05.100 --> 00:58:07.290 align:middle line:84% They say, you're a patient for our trial. 00:58:07.290 --> 00:58:09.570 align:middle line:90% Now let's genotype you. 00:58:09.570 --> 00:58:13.470 align:middle line:84% Because we knew that one out of six people was in Arg/Arg. 00:58:13.470 --> 00:58:17.440 align:middle line:84% If you were then met the criteria and were in Arg/Arg, 00:58:17.440 --> 00:58:19.080 align:middle line:90% you were enrolled in the trial. 00:58:19.080 --> 00:58:22.830 align:middle line:84% We then went through the people who we determined to be Gly/Gly 00:58:22.830 --> 00:58:26.760 align:middle line:84% and looked for a match, sex, lung function, and center 00:58:26.760 --> 00:58:29.290 align:middle line:90% where they were studying. 00:58:29.290 --> 00:58:33.565 align:middle line:84% So we ended up with these people. 00:58:33.565 --> 00:58:35.190 align:middle line:84% These were the baseline characteristics 00:58:35.190 --> 00:58:36.240 align:middle line:90% of the two groups. 00:58:36.240 --> 00:58:37.590 align:middle line:90% We matched on sex. 00:58:37.590 --> 00:58:41.040 align:middle line:84% And we came as close as you could. 00:58:41.040 --> 00:58:44.580 align:middle line:84% In terms of the proportions that were Caucasian, 00:58:44.580 --> 00:58:46.950 align:middle line:84% they worked out to be about the same. 00:58:46.950 --> 00:58:48.920 align:middle line:84% Their ages worked out to be the same. 00:58:48.920 --> 00:58:50.980 align:middle line:84% Their lung functions worked out to be the same. 00:58:50.980 --> 00:58:52.920 align:middle line:84% So we had two reasonably matched groups 00:58:52.920 --> 00:58:56.610 align:middle line:84% of people that differed with respect to their genotype 00:58:56.610 --> 00:59:00.070 align:middle line:84% at the beta-2 adrenergic receptor. 00:59:00.070 --> 00:59:03.900 align:middle line:90% So this was unexpected. 00:59:03.900 --> 00:59:05.790 align:middle line:84% What happened in this trial was different 00:59:05.790 --> 00:59:08.700 align:middle line:90% from our other trials. 00:59:08.700 --> 00:59:14.070 align:middle line:84% We had the patients on beta agonist 00:59:14.070 --> 00:59:15.930 align:middle line:90% on a regularly schedule basis. 00:59:15.930 --> 00:59:17.940 align:middle line:84% But they had a beta-agonist inhaler 00:59:17.940 --> 00:59:20.640 align:middle line:84% they were using if their asthma acted up. 00:59:20.640 --> 00:59:25.440 align:middle line:84% So when we did this trial, we switched their rescue inhaler 00:59:25.440 --> 00:59:28.110 align:middle line:84% from a beta-agonist inhaler to another type 00:59:28.110 --> 00:59:30.810 align:middle line:84% of rescue inhaler called Atrovent, an ipratropium 00:59:30.810 --> 00:59:34.380 align:middle line:84% bromide, which doesn't work at this receptor. 00:59:34.380 --> 00:59:36.510 align:middle line:84% And we allowed six weeks for patients 00:59:36.510 --> 00:59:39.480 align:middle line:84% to get used to using that treatment. 00:59:39.480 --> 00:59:43.120 align:middle line:84% Well, we hadn't expected this to happen. 00:59:43.120 --> 00:59:45.330 align:middle line:84% But what happened was in the Gly/Gly patients, 00:59:45.330 --> 00:59:48.180 align:middle line:90% their peak flow stayed fixed. 00:59:48.180 --> 00:59:49.870 align:middle line:84% Well, in the Arg/Arg patients who 00:59:49.870 --> 00:59:55.350 align:middle line:84% were using their albuterol four or five times a day, 00:59:55.350 --> 00:59:58.210 align:middle line:84% just kind of in regular use, when we switched them 00:59:58.210 --> 01:00:00.630 align:middle line:84% to ipratropium bromide over six weeks, 01:00:00.630 --> 01:00:03.750 align:middle line:84% it improved their lung function 30 liters a minute, 01:00:03.750 --> 01:00:06.150 align:middle line:90% actually 28 liters a minute. 01:00:06.150 --> 01:00:08.460 align:middle line:84% Now that's as good as you get out of most treatments 01:00:08.460 --> 01:00:10.300 align:middle line:90% you pay money for it. 01:00:10.300 --> 01:00:14.280 align:middle line:84% And all we did was switch them from albuterol 01:00:14.280 --> 01:00:16.230 align:middle line:90% to ipratropium bromide. 01:00:16.230 --> 01:00:19.350 align:middle line:84% But in our trial design, we said the zero point 01:00:19.350 --> 01:00:21.630 align:middle line:90% is going to be six weeks. 01:00:21.630 --> 01:00:24.240 align:middle line:84% So all our comparisons are made from here, 01:00:24.240 --> 01:00:28.140 align:middle line:84% where we had this big effect during the run-in, which 01:00:28.140 --> 01:00:31.200 align:middle line:90% we hadn't counted on. 01:00:31.200 --> 01:00:33.190 align:middle line:90% Then during the trial-- 01:00:33.190 --> 01:00:34.770 align:middle line:90% so the Arg/Arg is yellow. 01:00:34.770 --> 01:00:36.720 align:middle line:90% The Gly/Gly is blue-- 01:00:36.720 --> 01:00:40.350 align:middle line:84% They start off with peak flows of around 470. 01:00:40.350 --> 01:00:42.900 align:middle line:84% The Arg/Args get worse on active treatment 01:00:42.900 --> 01:00:45.720 align:middle line:90% while the Gly/Glys get better. 01:00:45.720 --> 01:00:49.020 align:middle line:84% While on placebo, the Arg/Args get better 01:00:49.020 --> 01:00:51.670 align:middle line:90% and the Gly/Glys get worse. 01:00:51.670 --> 01:00:54.330 align:middle line:84% This is the primary outcome peak flow. 01:00:54.330 --> 01:00:56.100 align:middle line:90% When we express it this way-- 01:00:56.100 --> 01:00:58.680 align:middle line:90% this is now morning peak flow. 01:00:58.680 --> 01:01:01.920 align:middle line:84% And so each of these bars represents a difference 01:01:01.920 --> 01:01:04.510 align:middle line:90% between active and placebo. 01:01:04.510 --> 01:01:07.660 align:middle line:84% So if active is worse than placebo, you go down. 01:01:07.660 --> 01:01:11.490 align:middle line:84% And if active is better than placebo, you go up. 01:01:11.490 --> 01:01:12.750 align:middle line:90% Arg is red. 01:01:12.750 --> 01:01:14.550 align:middle line:90% Gly is green. 01:01:14.550 --> 01:01:17.250 align:middle line:84% And you see that the difference between the two, 01:01:17.250 --> 01:01:19.260 align:middle line:84% highly statistically significant, 01:01:19.260 --> 01:01:23.760 align:middle line:84% with a value that's on the order of 23 liters per minute. 01:01:23.760 --> 01:01:27.750 align:middle line:84% We said 25 would have been clinically significant. 01:01:27.750 --> 01:01:31.920 align:middle line:84% It came close but didn't hit that bar. 01:01:31.920 --> 01:01:34.170 align:middle line:84% Then at evening peak flow, the difference 01:01:34.170 --> 01:01:36.330 align:middle line:90% is also in the same sign. 01:01:36.330 --> 01:01:39.060 align:middle line:84% But interestingly, there have been some recovery of lung 01:01:39.060 --> 01:01:41.520 align:middle line:90% function during the day. 01:01:41.520 --> 01:01:42.600 align:middle line:90% And our other outcomes. 01:01:42.600 --> 01:01:47.460 align:middle line:84% FEV1-- and difference again in the same sign. 01:01:47.460 --> 01:01:48.630 align:middle line:90% Arg/Args get worse. 01:01:48.630 --> 01:01:50.340 align:middle line:90% Gly/Glys get better. 01:01:50.340 --> 01:01:52.710 align:middle line:84% The difference is about 150 liters per minute. 01:01:52.710 --> 01:01:55.900 align:middle line:84% Everybody agrees this is statistically significant. 01:01:55.900 --> 01:01:56.970 align:middle line:90% When we look at symptoms. 01:01:56.970 --> 01:01:59.835 align:middle line:84% And here, a bigger number is more symptoms. 01:01:59.835 --> 01:02:01.710 align:middle line:84% There are more symptoms when the patients are 01:02:01.710 --> 01:02:03.540 align:middle line:84% using the drug on a regular basis 01:02:03.540 --> 01:02:05.710 align:middle line:84% than they are when they're using placebo. 01:02:05.710 --> 01:02:07.620 align:middle line:84% Again, in a sign of Gly/Glys doing 01:02:07.620 --> 01:02:10.110 align:middle line:84% better on active and Arg/Args doing 01:02:10.110 --> 01:02:13.980 align:middle line:84% worse on active treatment, so highly likely 01:02:13.980 --> 01:02:16.660 align:middle line:90% not to have occurred by chance. 01:02:16.660 --> 01:02:19.680 align:middle line:84% Now my final outcome here is how often the patients 01:02:19.680 --> 01:02:21.830 align:middle line:90% had to use a rescue inhaler. 01:02:21.830 --> 01:02:26.070 align:middle line:84% The ipratropium bromide was the non-beta-agonist rescue 01:02:26.070 --> 01:02:26.610 align:middle line:90% inhaler. 01:02:26.610 --> 01:02:29.940 align:middle line:84% Albuterol is a beta-agonist rescue inhaler. 01:02:29.940 --> 01:02:31.920 align:middle line:90% And we found the same thing. 01:02:31.920 --> 01:02:34.530 align:middle line:84% Remember, during this period, the patients 01:02:34.530 --> 01:02:38.790 align:middle line:84% are getting albuterol eight puffs a day already. 01:02:38.790 --> 01:02:40.950 align:middle line:90% And they're taking more of it. 01:02:40.950 --> 01:02:41.805 align:middle line:90% It's not enough. 01:02:41.805 --> 01:02:44.760 align:middle line:90% 01:02:44.760 --> 01:02:46.920 align:middle line:84% When the Gly/Glys are using at eight puffs a day, 01:02:46.920 --> 01:02:49.900 align:middle line:84% they cut back on their daily use. 01:02:49.900 --> 01:02:52.350 align:middle line:84% So the difference between these groups-- three puffs a day 01:02:52.350 --> 01:02:54.330 align:middle line:90% is highly significant. 01:02:54.330 --> 01:02:58.530 align:middle line:84% And in the sign that we're making the Arg/Arg patients 01:02:58.530 --> 01:03:02.320 align:middle line:90% worse with active treatment. 01:03:02.320 --> 01:03:05.760 align:middle line:84% So I think the clue out of this one is that about 1/6 01:03:05.760 --> 01:03:08.220 align:middle line:84% out of patients with asthma in the US 01:03:08.220 --> 01:03:12.460 align:middle line:84% are probably being made worse by their albuterol treatment, 01:03:12.460 --> 01:03:15.580 align:middle line:84% which is actually a very commonly used asthma therapy. 01:03:15.580 --> 01:03:20.670 align:middle line:84% Now this study was not calling for asthma exacerbations. 01:03:20.670 --> 01:03:23.490 align:middle line:84% To do that, we would have had to study, 01:03:23.490 --> 01:03:27.240 align:middle line:84% as I said, about 1,000 people for a year. 01:03:27.240 --> 01:03:29.850 align:middle line:84% And to study 1,000 people, you would've 01:03:29.850 --> 01:03:34.320 align:middle line:84% had to screen 3,000 or maybe more like 4,000, 01:03:34.320 --> 01:03:37.110 align:middle line:84% because it's going to be genotype stratified. 01:03:37.110 --> 01:03:38.730 align:middle line:90% But these two things track. 01:03:38.730 --> 01:03:42.000 align:middle line:84% And those kinds of trials are under consideration 01:03:42.000 --> 01:03:45.340 align:middle line:90% but are not completed yet. 01:03:45.340 --> 01:03:47.400 align:middle line:84% So I've given you three examples-- 01:03:47.400 --> 01:03:51.150 align:middle line:84% antileukotrienes where we've shown an effect based 01:03:51.150 --> 01:03:54.300 align:middle line:84% on what we know about the biology of the pathway but one 01:03:54.300 --> 01:03:57.720 align:middle line:84% of little pharmacoeconomic consequences, 01:03:57.720 --> 01:03:59.340 align:middle line:84% inhaled steroids where we weren't 01:03:59.340 --> 01:04:03.060 align:middle line:84% quite able to replicate finding [? three ?] genes. 01:04:03.060 --> 01:04:07.710 align:middle line:84% Well, we have a gene but probably not a SNP 01:04:07.710 --> 01:04:09.810 align:middle line:90% related to the steroid response. 01:04:09.810 --> 01:04:13.440 align:middle line:84% It'd also useful to try to do a controlled perspective 01:04:13.440 --> 01:04:15.510 align:middle line:90% trial that way. 01:04:15.510 --> 01:04:17.460 align:middle line:84% And then with the beta agonist, we've 01:04:17.460 --> 01:04:19.620 align:middle line:84% done the controlled perspective trial 01:04:19.620 --> 01:04:22.470 align:middle line:84% showing the genotype makes a difference in treatment 01:04:22.470 --> 01:04:23.110 align:middle line:90% response. 01:04:23.110 --> 01:04:26.680 align:middle line:84% So that it's more than just an idea, 01:04:26.680 --> 01:04:30.730 align:middle line:84% it's something we can actually reduce to practice. 01:04:30.730 --> 01:04:33.730 align:middle line:84% And so these are three examples in asthma treatment 01:04:33.730 --> 01:04:36.750 align:middle line:84% which tested the advantage of being a recurrent disease 01:04:36.750 --> 01:04:38.820 align:middle line:90% requiring chronic therapy. 01:04:38.820 --> 01:04:43.710 align:middle line:84% And if I were doing a trial where 01:04:43.710 --> 01:04:47.980 align:middle line:84% I wasn't allowed to take patients off treatment, 01:04:47.980 --> 01:04:49.980 align:middle line:84% right where the outcomes are going 01:04:49.980 --> 01:04:53.700 align:middle line:84% to be strokes, or heart, attacks or terrible events, then 01:04:53.700 --> 01:04:56.250 align:middle line:84% you'd be-- not have this kind of freedom. 01:04:56.250 --> 01:04:58.650 align:middle line:84% Even when you're using blood pressure as an outcome. 01:04:58.650 --> 01:05:01.275 align:middle line:84% You don't want to let someone's blood pressure go uncontrolled. 01:05:01.275 --> 01:05:04.140 align:middle line:84% But here, you can really show a genotype by treatment 01:05:04.140 --> 01:05:06.870 align:middle line:84% interaction different kinds of ways. 01:05:06.870 --> 01:05:09.630 align:middle line:84% And you can expand this to other kinds of diseases, 01:05:09.630 --> 01:05:11.580 align:middle line:84% doing the right kinds of treatment trials 01:05:11.580 --> 01:05:16.440 align:middle line:84% where you begin to enroll people by genotype and informative 01:05:16.440 --> 01:05:22.590 align:middle line:84% genotypes, which you find out from the evaluation of all data 01:05:22.590 --> 01:05:26.980 align:middle line:84% to decide whether this makes sense to move forward. 01:05:26.980 --> 01:05:29.520 align:middle line:84% So this is where I see, to me, one 01:05:29.520 --> 01:05:35.190 align:middle line:84% of the major advantages of genetics in actual medicine 01:05:35.190 --> 01:05:36.210 align:middle line:90% years. 01:05:36.210 --> 01:05:38.940 align:middle line:84% I mean, I can find the genes that cause things. 01:05:38.940 --> 01:05:41.340 align:middle line:84% But all I can do is give you bad or good news. 01:05:41.340 --> 01:05:44.100 align:middle line:84% You don't have a gene for Huntington's disease. 01:05:44.100 --> 01:05:47.082 align:middle line:84% You don't have the gene for Parkinson's disease or you do. 01:05:47.082 --> 01:05:48.540 align:middle line:84% Now you're going to worry about it, 01:05:48.540 --> 01:05:51.218 align:middle line:84% but you can't do anything about it. 01:05:51.218 --> 01:05:52.760 align:middle line:84% I might be able to tell you, you have 01:05:52.760 --> 01:05:55.680 align:middle line:90% a gene for a form of emphysema. 01:05:55.680 --> 01:05:58.530 align:middle line:90% And you shouldn't smoke. 01:05:58.530 --> 01:06:01.132 align:middle line:84% But still, there's very little the information 01:06:01.132 --> 01:06:03.090 align:middle line:84% that I can do with the information about a gene 01:06:03.090 --> 01:06:06.150 align:middle line:84% that I harbor or don't harbor, other than making decisions 01:06:06.150 --> 01:06:10.470 align:middle line:84% about my kids, and when I want to have kids, 01:06:10.470 --> 01:06:12.360 align:middle line:90% and who I want to marry. 01:06:12.360 --> 01:06:17.460 align:middle line:84% And it's interesting, in some regions of New York City where 01:06:17.460 --> 01:06:21.480 align:middle line:84% they have these arranged Ashkenazi Jewish marriages, 01:06:21.480 --> 01:06:23.130 align:middle line:84% they have a number of genetic diseases. 01:06:23.130 --> 01:06:27.150 align:middle line:84% They're actually genotyping the people 01:06:27.150 --> 01:06:29.730 align:middle line:84% in the arranged marriages so that they don't end up 01:06:29.730 --> 01:06:34.740 align:middle line:84% with Tay-Sachs disease and other commonly inherited-- 01:06:34.740 --> 01:06:36.840 align:middle line:90% but they're arranged marriages. 01:06:36.840 --> 01:06:40.560 align:middle line:84% A lot of marriages these days aren't arranged. 01:06:40.560 --> 01:06:42.420 align:middle line:84% So it's not terribly useful information. 01:06:42.420 --> 01:06:44.970 align:middle line:84% If on the other hand you have a disease, 01:06:44.970 --> 01:06:50.430 align:middle line:84% and I can look at your genotype and modify my treatment so that 01:06:50.430 --> 01:06:52.170 align:middle line:84% in one case, you're likely to get better. 01:06:52.170 --> 01:06:54.210 align:middle line:84% In another case, I know that a drug 01:06:54.210 --> 01:06:55.650 align:middle line:90% isn't going to work for you. 01:06:55.650 --> 01:06:59.850 align:middle line:84% That's a much more medically useful thing to know. 01:06:59.850 --> 01:07:02.070 align:middle line:84% I'm sorry that you inherited these genes that 01:07:02.070 --> 01:07:04.270 align:middle line:84% are going to cause you to have this bad disease. 01:07:04.270 --> 01:07:07.080 align:middle line:84% But at least I know from your genetic profile 01:07:07.080 --> 01:07:10.870 align:middle line:84% what treatment you're more likely to respond to it. 01:07:10.870 --> 01:07:13.470 align:middle line:90% And so that's why I think-- 01:07:13.470 --> 01:07:15.420 align:middle line:84% what is going to make a difference in genetics 01:07:15.420 --> 01:07:16.152 align:middle line:90% and medicine? 01:07:16.152 --> 01:07:17.610 align:middle line:84% This is going to make a difference. 01:07:17.610 --> 01:07:21.840 align:middle line:84% And we're seeing it now with leukemias. 01:07:21.840 --> 01:07:24.090 align:middle line:84% We're going through and doing these studies looking 01:07:24.090 --> 01:07:26.910 align:middle line:84% at patients that have a certain kind of leukemia who 01:07:26.910 --> 01:07:29.010 align:middle line:84% are likely to have a good treatment response 01:07:29.010 --> 01:07:32.425 align:middle line:84% and who aren't likely to have a good treatment response. 01:07:32.425 --> 01:07:33.800 align:middle line:84% What that means is that if you're 01:07:33.800 --> 01:07:36.720 align:middle line:84% sure in the good treatment response group, 01:07:36.720 --> 01:07:39.810 align:middle line:84% I can treat your leukemia and maybe not come 01:07:39.810 --> 01:07:41.910 align:middle line:90% so close to killing them. 01:07:41.910 --> 01:07:44.060 align:middle line:84% While if you're in a bad treatment response group, 01:07:44.060 --> 01:07:47.630 align:middle line:84% I'm going to get out everything I possibly can from day zero. 01:07:47.630 --> 01:07:49.890 align:middle line:84% And I'm going to try to wipe out every cell. 01:07:49.890 --> 01:07:51.890 align:middle line:84% So that I think this is much more 01:07:51.890 --> 01:07:55.760 align:middle line:84% medically useful than finding causative genes. 01:07:55.760 --> 01:07:58.790 align:middle line:84% We have to find genes that help us modify and understand 01:07:58.790 --> 01:08:00.420 align:middle line:90% our treatment response. 01:08:00.420 --> 01:08:03.500 align:middle line:90% So that's my story. 01:08:03.500 --> 01:08:04.000 align:middle line:90%