Tollers Take Two...

Yes, I know this is not a Toller...

Three and a bit weeks ago, I blogged about German breeder, Alexander Dauber, who had decided to outcross his Nova Scotia Tolling Retriever to an Australian Shepherd - the first step of a project aimed at introducing a shot of genetic diversity into the Toller.  Boy, from the response, you'd have thought I'd advocated murder ("You did... he's murdering the breed!), not just one man's effort to do what he thinks is right which no one else is remotely obliged to be involved with if they don't want to be.

It has become the blog's third most read post of all time. It attracted 74 comments that I published and another 10 or so that I thought were too awful to make public - although I'll share one of them now; this anonymous one from someone clearly very irate that I hadn't published Toller researcher Dr Danika Bannasch's reponse (which I had asked for and which she provided shortly after I'd published the post).

You promised to put Dr. Bannasch's responses on as soon as you got them. You've had them for weeks. Meantime, your anti-purebred shill "researcher" Maki has had multiple chances to comment, which have all been published. Put Dr. Bannasch's information up as you promised - or were you lying and now need time to "spin" things your way?

The rogue breeder has no proven info on health other than what he claims on his website. The dogs are not on OFA nor are they in Tollerdata, which responsible breeders use to post real results, and not just "my vet says" bs put out by people who can't be bothered to follow any rules.

Many people feel if haplotypes were the reason for him using an Aussie male (when Goldens were approved), then what possible excuse is there for not doing it ahead of time? ...except that he's no more than the usual Doodle breeder, sadly given blessing by someone whose never bred in their life.

Nice. And, of course, there's the small point that health test results for German dogs are not normally listed on the (American) OFA database.

I have, in fact, written to Dr Bannasch to apologise and explain the reason for the delay - which is that I have been waiting for a response from Dr Maria Wilbe, another Toller-owning scientist. I explained that I felt that the two responses would be better incorporated into a new post rather than editing into the original, as most people do not go back to review posts they have already read. I also felt they were too important to just post as a Comment where they might not be found.

Anyway, I received the response from Maria Wilbe two days ago and it transpires that a whole bevy of impressive names have signed up to it too, so I am finally in a position to return to the Tollers as promised.

First, if you haven't already read the original post, please do (you can find it here) before reading on.

To quickly recap for those who need a quick reminder, essentially, the fuss centres on  two issues. The first is whether Toller diversity is compromised sufficiently to warrant an outcross; and the second is over the usefulness of DLA haplotyping testing as a means to reduce the incidence of immune-mediated disease in the breed.

On the first, there is clearly a bit of a science stand-off (as documented in the first post), muddied at times by a sometimes near-hysterical response from some Toller breeders to the very idea of sullying their breed with foreign blood. Some won't even consider seeking out and using the unregistered Tollers that can still be found in Canada, one breeder dismissing them as just "Toller-like mutts".

Regarding DLA haplotype testing, this is a DNA test which looks at a key section of the immune system. In 2009, Swedish scientist Maria Wilbe and her co-authors found an association between one particular DLA haplotype* and immune-mediated rheumatic disease (IMRD). This is one of two troublesome immune-mediated conditions found in Tollers (and which together are often called Toller Disease). The team found an elevated risk of developing the condition in dogs with one copy of this haplotype - and a higher risk in dogs that had two copies of it. (Abstract here.)

(*haplotype = a group of genes inherited as one)

So, given that a test is now available that allows you to ensure that there would be no doubling up of the IMRD-risk haplotype in any puppies bred,  surely it would make sense for breeders to avail themselves of the test? Well, no - not according to the scientists who have signed the statement below. In fact, they absolutely insist that breeders should not make breeding decisions on the basis of any such test until all the genetic risk factors (and it is true that there are others) are known.

This statement and Dr Bannasch's are in full below (as well as Dr Bannasch's answers to three supplemental questions I asked).

As you'll see, Dr Bannasch views outcrossing as an absolute a last resort and argues that Tollers are only as inbred as any other pure breed. (Not a huge comfort, surely?) I was also disappointed that she feels it impossible to collect reliable information re the incidence of immune-mediated disorders in the breed given that it is perfectly feasible to recruit a birth cohort of Tollers and follow them through their lives. (This has been done in other breeds.) And I was surprised that Dr Bannasch cited laboratory mice as an example of a healthy inbred population when, in fact, over 90 per cent of these inbred lines die out in the process of creating them, many do indeed suffer from specific genetic problems and all have to be kept in a protected environment - unable to withstand the environmental, bacterial and viral onslaught they would be exposed to in the real world.  Not something we'd want for our dogs.

As for the joint statement issued by the various researchers involved in DLA testing, I do understand the scientists' caution, but I believe that if some early adopters want to expore DLA testing then they should be free to do so, as long as they are fully aware that it is only one small part of the story and things might not work out as planned. I would hope too, that, what they discover along the way is passed on to the scientists to help their research - and the breed.

Why wouldn't you want to avoid doubling up on a haplotype if its been found that two copies of it increases the risk of IMRD, even if other genetic risk factors are involved? After all, cutting down on butter is only one way of reducing cholesterol, but it is better than doing nothing at all. 

I can also see some sense in, generally, using DLA testing to try to ensure as diverse an immune system as possible in a breed as we know that is, in general, a good thing (notwithstanding that some individual haplotypes might be associated with a specific disease).  In fact , this is the whole basis of the commerial DLA haplotype testing offered by Professor Hannes Lohi at Genoscoper whose name appears on the statement below advising against DLA testing as a means of reducing the risk of immune-mediated disease in Tollers.

DLA testing also affords an opportunity to increase the number of dogs carrying very rare haplotypes. Although 11 in total have been found so far in the breed, only five have been found with any great frequency and the rest are vanishingly rare. This table is in Finnish but you'll get the idea..and see how common the IMRD risk haplotype is. (NB again, it is only a risk factor and one of several. There are Tollers who have one and two copies of this haplotype who have never gone on to develop IMRD and, equally, those that don't have any copies of it who have.)

Click to enlarge

So there we go. As clear as mud, I expect. But it's a complex subject.

DR BANNASCH'S RESPONSE

I would be the first person to support purposeful cross breeding and in the event that we discover through our research that it is the ONLY way to correct an inherited problem in the breed I will be the first one leading the effort.

The most important point that you and others are missing is that the Maki report does not compare the NSDTR to other purebred breeds. This could be done for all breeds and may uncover that certain breeds are in more critical shape than others.  The NSDTR breeders are fortunate that since the breed was more recently admitted to registration status we have computer access to complete pedigrees since that time.  That is what gives the apparently very high inbreeding coefficients.  If this same analysis could be done in other breeds the numbers would likely be similar. It is not correct to compare the numbers to wild animal populations and make conclusions about the dire straits that this breed is in.  You assume that tollers are much worse off than other breeds and I believe based on our unpublished data (see below) they are about average among purebreds.

My laboratory has been working on identifying the genes that causes tollers to develop Addison's disease.  As part of this work we have genotype data from the NSDTR as well as other breeds and can compare the genotype based inbreeding coefficient.  These numbers are based on data for about 30,000 markers in 10 unrelated animals.

Dachshund- 0.1242, Lab 0.1635, Beagle 0.2043, NSDTR 0.2064, German Shepherd 0.3150 and greyhound 0.3238

We also performed Y haplotype analysis of 33 tollers and obtained diversity values of 0.76 +/- 0.04 based on identifying 5 different haplotypes in the NSDTR.  You can see numbers from other breeds in the following paper for comparison:

Y chromosome haplotype analysis in purebred dogs (pdf) Danika L. Bannasch,1 Michael J. Bannasch,2* Jeanne R. Ryun,1* Thomas R. Famula,3 Niels C. Pedersen

Review

Inbreeding and genetic diversity in dogs: Results from DNA analysis  (Claire M. Wade)

You also state that 11 DLA haplotypes is "low" for a breed but actually although most of the scientific manuscripts on the subject do not have the haplotypes published in this way, 11 is above average for a purebred dog breed. Please see partial reference list below.

This paper found 11 haplotypes in the GSD:

MHC class II risk haplotype associated with Canine chronic superficial keratitis in German Shepherd dogs. Päivi Jokinena, Elina M. Rusanenc,  Lorna J. Kennedy and Hannes Lohi.

This paper found 6 haplotypes in Pugs:

Necrotizing meningoencephalitis of Pug Dogs associates with dog leukocyte antigen class II and resembles acute variant forms of multiple sclerosis. K. A. Greer et al.

This paper found 9 haplotypes in Weimaraners:

Expanded dog leukocyte antigen (DLA) single nucleotide polymorphism (SNP) genotyping reveals spurious class II associations. N. Safraa, N.C. Pedersena, Z. Wolfa, E.G. Johnsona, H.W. Liua, A.M. Hughes, A. Young and D.L. Bannasch

This paper found 6 haploytpes in over 500 Dobermans:

Association of hypothyroid disease in Doberman Pinscher dogs with a rare major histocompatibility complex DLA class II haplotype. (pdf) L. J. Kennedy1, H. J. Huson, J. Leonard, J. M. Angles, L. E. Fox, J. W. Wojciechowski, C. Yuncker & G. M. Happ

JH supplemental questions:

JH: what is your estimate of the incidence of immune-mediated disorders in the breed? 
DB: Unfortunately this number is extremely difficult to get or even estimate.  When polls are taken more people with diseased dogs are likely to respond.  The ideal sampling would be to follow a cohort of dogs and report their incidence of disease.  Obviously this is impossible.

JH: an Ne of 18  is extremely low  - Calboli et al found only one breed lower of the 10 they looked at.  Why do you not consider this a problem? (Ne is "effective populaton size" - a measure of genetic diversity).
DB:Again if you look closely they followed out nine generations not to the beginning of the closed stud book.  It is like comparing oranges to steak.

JH: how is saying that other breeds are worse an argument for not taking action in the Toller?
DB: I don't know if you are familiar with inbred mice.  There are 100s of laboratory strains that are completely inbred- ie homozygous at every locus.  They breed prolifically and are healthy.  They are not living in the wild but neither are domestic dogs.  Purebred dogs are definitely inbred and have small effective population sizes.  The majority of them are healthy, look and act like each other (within a breed) which is what people want. It may be that there are some breeds that are in danger- I have not seen any scientific evidence for that.  The "action" that I have read about in the Toller was not based on a scientifically documented need.  In addition the "action" that I read included testing parents of the outcross to ensure that they did not have or carry all the diseases that NSDTR have been documented to get.  The problem with that is we do not have tests for those diseases...  Outcrossing without a clear scientific plan does not make sense to me.

STATEMENT FROM MARIA WILBE ET AL

To whom it may concern

A statement concerning the genetic basis for the immune-mediated rheumatic disease and steroid-responsive meningitis arteritis disease complex in Nova Scotia duck tolling retriever dogs

It has come to our attention that our recent publications concerning the genetic risk factors for the immune-mediated rheumatic disease (IMRD) and steroid-responsive meningitis arteritis (SRMA) disease complex in Nova Scotia duck tolling retriever dogs or “Tollers”, has left some uncertainty concerning the genetic basis for the disease complex. We would therefore like to explain our views and to clarify this subject and its implications for breeding. The disease is complex and has many genetic risk factors and therefore we cannot provide recommendations for breeders exclusively on the basis of genetic testing for dog leukocyte antigen (DLA) class II genotype.

1. The inheritance of the IMRD and SRMA disease complex All our data and all data we are aware of concerning the inheritance of IMRD and SRMA disease complex indicate complex inheritance. This means that there are several genetic factors involved that will influence the disease phenotype. Furthermore, environmental factors will also influence disease status. This means that any given Toller dog has inherited a particular combination of genetic risk factors and that the development of and severity of disease will depend on which combination of the genetic risk factors it has inherited, but also the overall genetic background and the environmental factors that the dog will experience during its lifetime. Importantly, like in other complex diseases, as a consequence of unique environmental exposure the disease aetiology for the disease complex will differ in different Tollers even though they have inherited the same set of genetic risk factors.

It should be noted that our genome-wide association (GWA) study presented in our Nature Genetics paper from 2010 (Wilbe et al. Nature Genetics 42:250-254) was based primarily on a case- control population of Tollers from Sweden and Finland and validated using Tollers from the US. In these Tollers, significant association was obtained for all five regions. Importantly, some were stronger than others and only two was shared between IMRD and SRMA. The frequency of the actual genetic risk factors may differ in different Toller populations present in different countries. However, they are likely to be similar given the population structure in Tollers and how the breed was created. All of the current Tollers, world-wide, are derived from a small group of dogs that survived two devastating outbreaks in the early 20th century of canine distemper virus epidemics. This resulted in what geneticists call “genetic bottle-necks” and a strong founder effect for the genetic risk factors causing IMRD and SRMA.

The results from the GWA study identified five regions containing many strong candidate genes involved in T-cell activation (Wilbe et al. Nature Genetics 42:250-254). In our Immunogenetics paper from 2009 (Wilbe et al. Immunogenetics 61:557-564) we showed that dog leukocyte antigen (DLA) is another genetic risk factor for development of IMRD. The studies provided conclusive evidence that there are multiple genetic risk factors underlying the IMRD and SRMA disease complex. Importantly, we showed that some of these risk factors were specific for IMRD and that some were common between IMRD and SRMA. The actual mutations causing the disease have not yet been conclusively determined. Intensive research efforts in our laboratory are in progress to identify and validate such mutations. When the mutations have been identified and correlated genetically to disease development genetic tests for all the mutations can be developed.

2. The role of dog leukocyte antigens (DLA) class II in Immune-mediated rheumatic disease The results presented in our Immunogenetics paper from 2009 (Wilbe et al. Immunogenetics 61:557- 564) identified one DLA class II type as a genetic risk factor for the immune-mediated rheumatic disease (IMRD) but not for steroid-responsive meningitis arteritis (SRMA). If a dog has inherited the risk DLA class II type from both parents it likely has an increased risk of developing IMRD. Importantly, the DLA class II type is not the only genetic risk factor, which means that some dogs without this risk factor can still develop IMRD and the opposite is also true. Some individuals with the DLA class II risk type do not develop disease. Our data clearly showed that homozygosity for the risk DLA type is increased among IMRD-affected Tollers. The results from this study were based on a case-control population, which means that we used all our cases and compared to the same number of healthy dogs. Therefore, neither the frequencies of diseased dogs nor the frequencies of haplotypes do reflect the total frequency in the Toller population. A total of five DLA class II haplotypes was identified and this is similar to most other dog breeds. Furthermore, Hughes et al. (Tissue Antigens 75(6):684-90, 2010) identified two additional haplotypes which gives the Tollers a total of seven known DLA haplotypes. Typically, one or a few DLA types are increased in frequency in any given dog breed. However, we cannot accomplish reduced incidence of IMRD only based on a breeding practice based on DLA genotyping. Inadvertent increase in the frequency of any of the other five known genetic risk factors may be a consequence. We anticipate that when we have DNA tests for all genetic risk factors and knowledge of how they interact we may be able to give potential breeding recommendations how to reduce incidence of the disease.

3. DNA tests and Recommendations for breeding Commercially available DNA tests for DLA have been offered to Toller breeders since 2010. This DNA test can be used to identify carriers of the DLA risk type in heterozygous or homozygous form. Any DNA laboratory skilled in the art of DNA testing can perform this test and there is no patent protecting its use. Importantly, we have not yet established diagnostic DNA tests for the other five genetic risk factors. The establishment of such tests will require some further research. However, at present testing for DLA only is of limited use. We cannot provide recommendations for breeders exclusively on the basis of genetic testing for dog leukocyte antigen (DLA) class II genotype. We strongly discourage breeders to perform their dog breeding only on the basis of DLA genotyping. This may lead to increased risk of inheriting unwanted combination of other major genetic risk factors for the disease complex. Attempts to reduce the incidence of the IMRD and SRMA disease complex can and should only be based on genotype data on all the genetic risk factors, thereby avoiding the most disadvantageous combinations of genetic risk factors. Therefore, there is no current way for breeders to perform DNA testing to reduce or eliminate this disease.

Göran Andersson1, Dannika Bannasch2, Helene Hansson-Hamlin3, Kerstin Lindblad-Toh4,5, Hannes Lohi6, Claire Wade7 and Maria Wilbe1 1Department of Animal Breeding and Genetics, Swedish University of Agricultural Sciences (SLU), Biomedical centre, Box 597, SE-751 24 Uppsala, Sweden.
2Department of Population Health and Reproduction, School of Veterinary Medicine, University of California Davis, Davis, CA 3Department of Clinical Sciences, SLU, Box 7054, SE-750 07 Uppsala, Sweden 4Department of Medical Biochemistry and Microbiology, Uppsala University, Box 597, SE-751 24 Uppsala, Sweden.
5Broad Institute of Harvard and MIT, 7 Cambridge Center, Cambridge, Massachusetts 02142, USA. 6Department of Veterinary Biosciences, University of Helsinki, Box 63, 00014 Helsinki, Finland. 7Faculty of Veterinary Science, University of Sydney, NSW 2006, Australia
September 12, 2011

Edit September 20: added links to Dr Bannasch's references