Nutrition and the Skeletal Health of Dogs and Cats

Nutrition and the skeletal health of dogs and cats

Lay-out: Proefschrift-aio.nl Cover: Proefschrift-aio.nl

ISBN: 978-90-393-6174-0 Nutrition and the skeletal health of dogs and cats

De invloed van voeding op het skelet van honden en katten (met een samenvatting in het Nederlands)

Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op donderdag 28 augustus 2014 des middags te 4.15 uur

door

Ronald Jan Corbee

geboren op 28 april 1979 te Alkmaar Promotoren: Prof.dr. H.A.W. Hazewinkel Prof.dr. J.W. Hesselink Copromotoren: Dr. M.A. Tryfonidou Dr. A.B. Vaandrager Contents

Chapter 1: General introduction, aim and scope of the study 7

Part I: Effects of nutrition on osteoarthritis

Chapter 2: Obesity and osteoarthritis 15 Chapter 3: Obesity in show dogs 33 Chapter 4: Obesity in show cats 49 Chapter 5: The effect of dietary long-chain omega-3 fatty acids 61 supplementation on owner’s perception of behavior and locomotion in cats with naturally occurring osteoarthritis

Part II: Vitamin A and D requirements in relation to skeletal health

Chapter 6: The interaction of vitamin A and vitamin D with emphasis 79 on bone metabolism Chapter 7: Cutaneous vitamin D synthesis in carnivorous species 103 Chapter 8: Composition and use of puppy milk replacers in German 115 Shepherd puppies in the Netherlands Chapter 9: Dietary vitamin D supplementation during early growth 129 does not protect against medial coronoid disease in Labradors Chapter 10: Vitamin D status before and after hypophysectomy in dogs 149 with pituitary-dependent hypercortisolism Chapter 11: Ground reaction forces of walking cats; assessment and 163 comparison with walking dogs Chapter 12: Chronic vitamin A supplementation in cats results in 185 minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation Chapter 13: General discussion 205

Summary 219 Samenvatting 225 Curriculum vitae and list of publications 231 Acknowledgements 235 6 Chapter 1

General introduction, aim and scope of the study

7 1

8 General introduction, aim and scope of the study

Nutrition is about balance; nutritional imbalances can cause several pathological 1 changes in companion animals. In this thesis, the influence of nutrition on skeletal health is studied. Nutrition supplies building blocks for bone (e.g. calcium, phosphate) and provides essential co-factors for enzymes (e.g. copper and manganese) which influence bone cells to either produce (osteoblasts) or resorb (osteoclasts) extracellular matrix with the aid of vitamin A and vitamin D. In part I of this thesis the possible effects of nutrition on osteoarthritis, like high caloric intake and supplementation of long chain polyunsaturated fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), are studied. In Part II of this thesis, the effects of imbalances of vitamin A and D on skeletal health are studied.

Part I

Obesity is a common disease in companion animal practice. It is a condition in which excess body fat has accumulated to the extent that it may be harmful to health, caused by an imbalance between energy intake and energy expenditure. In dogs and cats obesity often exacerbates osteoarthritis (OA). To elucidate the mechanisms by which OA and obesity are linked, a review of the literature will be performed (Chapter 2). Several dog and cat breeds are suggested to have an increased prevalence of obesity; however this is not substantially clarified by current scientific evidence. The aim of the study as described in Chapter 3 is to identify the dog breeds with an increased risk for obesity. The same for cat breeds in Chapter 4. As these studies are performed in show dogs and show cats, respectively, these studies also evaluate if what owners and judges regard as normal or beautiful does correspond with the optimal body condition.

In dogs, OA is a common disease and several treatment options have been investigated and are generally accepted by clinicians. Recently, the high prevalence of OA in cats was identified (Slingerland et al. 2011), and it was clear that these cats demonstrate more subtle behavioral changes, rather than obvious lameness as can be observed in dogs suffering from OA. One of the nutraceuticals used in dogs suffering from OA are dietary long-chain omega-3 fatty acids EPA and DHA. EPA and DHA are the building blocks for prostaglandins and leukotrienes with anti-inflammatory effects. These free fatty acids have been proved to be effective in dogs suffering from OA by improving locomotion, as was demonstrated by force plate analysis (Roush et al. 2010). Furthermore, supplementation of EPA and DHA was effective by lowering the dosage of non-steroid anti-inflammatory drugs necessary to control lameness in dogs with OA (Fritsch et al. 2010). There is only little evidence for effects of EPA and DHA in cats with OA, as only one study measured the effects of a therapeutic diet (including EPA and DHA) on lameness in cats (Lascelles et al. 2010). The aim of the study described in Chapter 5 is to investigate if supplementation of EPA and DHA will have beneficial effects on behavior and locomotion in client-owned cats suffering from OA.

9 1 Part II

Omnivore and herbivore species are able to synthesize vitamin D from 7-dehydrocholesterol in the skin under influence of UVB-light. Dogs and cats are unable to synthesize sufficient amounts of vitamin D in the skin, unlike rats (How et al. 1994). This is possibly caused by an increased activity of delta-7-dehydrocholesterol reductase (d7DHCr), as was demonstrated in cats who were able to synthesize sufficient amounts of vitamin D in the skin when they received an inhibitor of d7DHCr (Morris 1999). The aim of Chapter 7 is to identify if other carnivorous species than the canine and feline species are also unable to synthesize significant amounts of vitamin D in the skin.

Nutrients influence bone metabolism, and thus the development of orthopedic diseases. Several studies demonstrate the detrimental effect of high and low calcium intake, as well as high vitamin D intake, on skeletal development in dogs (Hazewinkel et al. 1991, Schoenmakers et al. 2000, Tryfonidou et al. 2002). Vitamin D influences the process of mineralization of newly formed osteoid and newly formed cartilage, and thus of endochondral ossification during skeletal growth. It has also been demonstrated that nutritional imbalance during a short period early in life can result in developmental orthopedic diseases (Hazewinkel and Schoenmakers 2006, Tryfonidou et al. 2003). Given that puppies are often fed with milk replacers for several reasons, the aim of the study in Chapter 8 is to identify if the vitamin D concentration in puppy milk replacers (PMRs) is similar to the vitamin D concentration of bitch milk and whether the use of PMR poses a risk factor for orthopedic diseases later in life. One of the common developmental orthopedic diseases in young dogs is medial coronoid disease (MCD). A recently published study demonstrated that MCD is caused by delayed endochondral ossification (Lau et al. 2013, Thompson 2007), which might be due to a (relative) vitamin D deficiency. The aim of the study in Chapter 9 is to identify if supplementation of the daily ration with extra vitamin D during the first growth phase (i.e., 3-18 weeks of age) can prevent development of MCD by stimulating endochondral ossification, including terminal differentiation of chondrocytes, and mineralization of the cartilaginous template of the developing medial coronoid process.

In adult dogs with Cushings disease (hyperadrenocorticism) an important source of GH and consequently of IGF-I is suppressed by the negative feedback of cortisol on the pituitary gland. In 85% of the cases of Cushings disease, hyperadrenocorticism is pituitary dependent. Treatment options are either destruction of the adrenal glands, medication to reduce the release of cortisol, or removal of the pituitary gland (e.g. hypophysectomy). The removal of the pituitary gland results in the removal of an important source of GH, and in humans this results in osteoporosis due to decreased 1,25-dihydroxyvitamin D plasma levels and concomitant increased parathyroid hormone (PTH) secretion. The increase of PTH secretion in dogs has been associated with decreased quality of life, appetite, activity, strength, and life span (Nagode et al. 1996). The aim of Chapter 10 is to determine whether in dogs with pituitary dependent hyperadrenocorticism

10 General introduction, aim and scope of the study

the PTH-vitamin D axis is disturbed, and whether hypophysectomy leads to lower GH 1 and thus IGF-I blood levels resulting in lowered 1,25-dihydroxyvitamin D synthesis and consequently hyperparathyroidism. The study was to elucidate the necessity of supplementation of vitamin D metabolites to this significant patient group with pituitary dependent hyperadrenocorticism before and/or after hypophysectomy.

In human medicine, hypervitaminosis A is associated with liver fibrosis (Castaño et al. 2006), while in cats excessive consumption of vitamin A is thought to cause hyperostosis, mainly located at the cervical vertebrae (Polizopoulou et al. 2005). The mechanisms, by which this hyperostosis develops is still unknown. The role of vitamin A as the sole responsible nutrient was already questioned following a study in which adult cats were fed a diet high in vitamin A for 2 years without demonstrating any effects on the skeleton (Freytag et al. 2003). Notably, the clinical cases describing hyperostosis due to hypervitaminosis A, always involve consumption of raw liver, which contains high amounts of vitamin A and increased levels of vitamin D. Therefore, in part II of this thesis we aim to identify if hypervitaminosis A has to be accompanied by hypervitaminosis D in order to develop the typical signs of hyperostosis in cats. In order to do so, the possible interactions between vitamin A and vitamin D metabolism will be identified by reviewing the literature (Chapter 6). Furthermore, in order to have more objective outcome parameters of locomotion in in vivo studies with cats described in part II of the thesis, we decided to develop and validate force plate analysis (FPA) in cats (Chapter 11). The clinical and radiological findings, the force plate analysis, and pathology of liver biopsies in cats receiving either high amounts of vitamin A, or high amounts of vitamin A and of vitamin D are described in Chapter 12. The results of the reviews and of the studies described in this thesis will be discussed in Chapter 13.

11 1 References • Lau, S.F., Hazewinkel, H.A.W., Grinwis, G.C.M., Wolschrijn, C.F., Siebelt, M., Vernooij, J.C.M., • Castaño, G., Etchart, C., Sookoian, S., 2006. Voorhout, G., Tryfonidou, M.A., 2013. Delayed Vitamin A toxicity in a physical culturist endochondral ossification in early medial patient: A case report and review of the coronoid disease (MCD): A morphological and literature. Annals of Hepatology 5, 293-295. immunohistochemical evaluation in growing • Freytag, T.L., Liu, S.M., Rogers, Q.R., Morris, J.G., Labrador retrievers. The Veterinary Journal 2003. Teratogenic effects of chronic ingestion 197, 731-738. of high levels of vitamin A in cats. Journal • Morris, J.G., 1999. Ineffective vitamin D of Animal Physiology and Animal Nutrition synthesis in cats is reversed by an inhibitor of 87, 42-51. 7-dehydrocholestrol-δ7-reductase. Journal of • Fritsch, D.A., Allen, T.A., Dodd, C.E., Jewell, Nutrition 129, 903-908. D.E., Sixby, K.A., Leventhal, P.S., Brejda, J., • Nagode, L.A., Chew, D.J., Podell, M., Hahn, K.A., 2010. A multicenter study of the 1996. Benefits of calcitriol therapy and effect of dietary supplementation with fish serum phosphorus control in dogs and oil omega-3 fatty acids on carprofen dosage cats with chronic renal failure. Both are in dogs with osteoarthritis. Journal of the essential to prevent or suppress toxic American Veterinary Medical Association hyperparathyroidism. Veterinary Clinics of 236, 535-539. North America – Small Animal Practice 26, • Hazewinkel, H.A.W., Schoenmakers, I., 2006. 1293-1330. Hormonal and skeletal effects of excessive • Polizopoulou, Z.S., Kazakos, G., Patsikas, M.N., calcium intake during partial weaning and Roubies, N., 2005. Hypervitaminosis A in the at prepubertal age in dogs. Compendium cat: a case report and review of the literature. on Continuing Education for the Practicing Journal of Feline Medicine and Surgery 7, Veterinarian 28, 55. 363-368. • Hazewinkel, H.A.W., Van den Brom, W.E., Van ‘T • Roush, J.K., Cross, A.R., Renberg, W.C., Dodd, Klooster, A.T., Voorhout, G., Van Wees, A., 1991. C.E., Sixby, K.A., Fritsch, D.A., Allen, T.A., Jewell, Calcium metabolism in Great Dane dogs fed D.E., Richardson, D.C., Leventhal, P.S., Hahn, diets with various calcium and phosphorus K.A., 2010. Evaluation of the effects of dietary levels. Journal of Nutrition 121, S99-106. supplementation with fish oil omega-3 • How, K.L., Hazewinkel, H.A.W., Mol, J.A., 1994. fatty acids on weight bearing in dogs with Dietary vitamin D dependence of cat and osteoarthritis. Journal of the American dog due to inadequate cutaneous synthesis Veterinary Medical Association 236, 67-73. of vitamin D. General and Comparative • Schoenmakers, I., Hazewinkel, H.A.W., Endocrinology 96, 12-18. Voorhout, G., Carlson, C.S., Richardson, D., • Lascelles, B.D.X., DePuy, V., Thomson, A., 2000. Effect of diets with different calcium Hansen, B., Marcellin-Little, D.J., Biourge, V., and phosphorus contents on the skeletal Bauer, J.E., 2010. Evaluation of a therapeutic development and blood chemistry of growing diet for Feline degenerative joint disease. great danes. Veterinary Record 147, 652-660. Journal of Veterinary Internal Medicine • Slingerland, L.I., Hazewinkel, H.A.W., Meij, B.P., 24, 487-495. Picavet, P., Voorhout, G., 2011. Cross-sectional study of the prevalence and clinical features of osteoarthritis in 100 cats. The Veterinary Journal 187, 304-309.

12 General introduction, aim and scope of the study

• Thompson, K., 2007. Bones and Joints. In: 1 Jubb, Kennedy & Plamer’s Pathology of Domestic Animals (5th Ed), Editor: M. Grant Maxie, p. 140. • Tryfonidou, M.A., Holl, M.S., Stevenhagen, J.J., Buurman, C.J., Deluca, H.F., Oosterlaken- Dijksterhuis, M.A., Van Den Brom, W.E., Van Leeuwen, J.P.T.M., Hazewinkel, H.A.W., 2003. Dietary 135-fold cholecalciferol supplementation severely disturbs the endochondral ossification in growing dogs. Domestic Animal Endocrinology 24, 265-285. • Tryfonidou, M.A., Stevenhagen, J.J., Van Den Bemd, G.J.C.M., Oosterlaken-Dijksterhuis, M.A., Deluca, H.F., Mol, J.A., Van Den Brom, W.E., Van Leeuwen, J.P.T.M., Hazewinkel, H.A.W., 2002. Moderate cholecalciferol supplementation depresses intestinal calcium absorption in growing dogs. Journal of Nutrition 132, 2644-2650.

13 Part I: Effects of nutrition on osteoarthritis

14 Chapter 2

Obesity and osteoarthritis

Parts of this manuscript have been published: Herman A.W. Hazewinkel Ronald J. Corbee

Obesity and Osteoarthritis: Causes and Management. Proceedings of the 2011 Nestlé Purina Companion Animal Nutrition Summit; Tucson, AZ, USA

15 Part I

Abstract

2 Obesity in several species is associated with an increased risk for osteoarthritis (OA). Besides the increased weight bearing of the joints in obese individuals, other mechanisms play an important role. Leptin and pro-inflammatory cytokines derived from adipose tissue perpetuate the cycle of OA, causing more pain and damage of articular cartilage. In young individuals overweight conditions should be prevented, as this might influence the set-point of leptin later in life. Individuals already suffering from OA are more prone to overweight conditions due to decreased energy expenditure by decreased activity and therefore, prevention of overweight conditions is important. Improvement of OA-associated pain and pathology has been noted during weight loss. Therefore, prevention and treatment of overweight conditions are the most important aspects in OA prevention and management.

16 Obesity and osteoarthritis

Introduction

In the western world of abundance, underweight conditions are rare. In contrast, 2 overweight conditions are getting more and more common in humans and pets (Vis and Hylands 2013, German 2006), and have become the new standards for our house dogs and cats (Becker et al. 2012, Courcier et al. 2011), which has consequences for their health and wellbeing. The most common cause of overweight conditions is energy intake exceeding energy expenditure. Several risk factors for overweight conditions in dogs and cats have been identified, such as breed/genetics (Jeusette et al. 2010), neuter status (Laflamme 2005), orthopedic diseases (and thus decreased activity) (McGreevy et al. 2005), type of diet (Kienzle et al. 1998), change of lifestyle (Robertson 2003), body condition score of the owner (Kienzle et al. 1998, Nijland et al. 2010), and the socioeconomic circumstances of the owner (McGreevy et al. 2005, Robertson 2003).

Because of improved knowledge and development of new therapeutic possibilities, people and pets live longer (Salomon et al. 2013). This increasing age is coincided by an increased prevalence of osteoarthritis (OA) in people (Sulzbacher 2013), dogs (Huck et al. 2009, Marshall et al. 2009), and cats (Slingerland et al. 2012).

The association between overweight conditions during life and development of OA has been demonstrated in dogs (Kealy et al. 2002), and people (Wang et al. 2013). The aim of this review is to elucidate the mechanisms by which overweight conditions influence OA development and OA associated conditions, as well as evaluating the effects of weight loss on OA associated pain and lameness.

Obesity

Obesity (>20% excess weight) and overweight (>10% excess weight) conditions have a growing incidence in humans, as well as in dogs and cats (Vis and Hylands 2013, German 2006). The sedentary life style of people, with easy access to energy dense foods, contributes to that (Racette et al. 2003). During evolution, the people and animals that were best adapted to store fat had more chance of survival in scarce periods; therefore there was a selection in favor of “easy keepers” (Speakman 2013). Nowadays, people and pets in the western world live in an environment that is completely different, with a constant abundance of food. This is a major risk factor for obesity, as people and pets are unable to cope with these new conditions. Another explanation might be a failure of adaptation/activation of brown fat tissue to the new (warm/indoor) housing conditions (Rothwell and Stock 1979). Normal weight individuals respond well to satiating signals, which prevent them from becoming obese or overweight. Those satiating signals are stretch-receptors in the stomach and the intestine, gastrin, insulin/glucagon ratio, peptide YY, cholecystokinin (CCK), estrogen, dopamine, leptin, and many more (Begg and Woods 2013). For most of these satiating signals it takes some time during or after eating before they are becoming active at the satiety center in the ventromedial nucleus of the

17 Part I

hypothalamus. Therefore, eating rapidly predisposes for overweight conditions, which are often seen in humans in the western world (Oda-Montecinos et al. 2013). Dogs, by 2 nature, have an eating pattern that is similar to wolves and consists of few large meals and long periods of fasting, which is the case in some dog breeds; other dog breeds, just like feral dogs, are scavenging (Bradshaw 2006). Dogs are therefore not easily satisfied and are always eager for food and are eating rapidly (Houpt 1991). Feeding dogs ad libitum is therefore a risk factor for obesity (Kienzle et al. 1998, Grant et al. 2011), especially after neutering (Jeusette et al. 2004).

Cats regulate their feed intake better compared to dogs, as cats are used to eat multiple small meals in a quiet environment (Bradshaw 2006). Neutering in cats is a common procedure, which predisposes house cats to obesity (Colliard et al. 2009). Neutered animals have decreased estrogen levels resulting in decreased spontaneous activity, and as a result, decreased muscle mass and strength and decreased energy expenditure (Alexander et al. 2011, Brown 2008). Furthermore, decreased estrogen levels are responsible for a decreased sensitivity for CCK-dependent satiety signals (Geary et al. 2001) and decreased leptin receptor expression in the hypothalamus and in adipose tissue resulting in leptin insensitivity (Meli et al. 2004). Supplementation of estrogen normalized food intake in neutered cats, demonstrating the essential role of estrogen in satiety regulation (Cave et al. 2007). Leptin is an important satiating signal and is produced by white adipose tissue when fat is being stored. Leptin is regarded as the “adipostat” (represses food intake and promotes energy expenditure) in humans and mice (Feng et al. 2013), and probably also in dogs and cats. Hypogonadal men and women have significantly higher leptin levels compared to normal individuals indicating that they store more fat before being satiated (Rosenbaum and Leibel 1999). In obese animals, leptin levels are constantly high, with loss of peripheral receptor sensitivity, demanding higher leptin levels to induce satiety. Veterinarians should give good nutritional recommendations after neutering to prevent overweight conditions because of a less well functioning of the “adipostat”.

When an individual gets overweight, it becomes less sensitive for the satiating signals because of the chronic high leptin levels resulting in leptin insensitivity, thus increasing energy intake, which is an alarming situation and has been demonstrated in humans (Feng et al. 2013). Overweight animals have increased leptin levels (Ishioka et al. 2002) and this results in an increased sense of hunger when these animals have central leptin insensitivity, which results in the typical begging behavior. Ultimately, chronic supraphysiological leptin levels result in altered energy expenditure, as leptin also negatively affects the release of thyroid stimulating hormone (Feng et al. 2013). High birth weight due to maternal obesity or diabetes, inappropriate early post-natal nutrition, and rapid catch-up growth may also sensitize to increased risk of obesity in young animals due to an altered leptin set-point (Breton 2013). The increased risk for obesity was also demonstrated in men in case of maternal famine (Ravelli et al. 1999). Maternal feeding status influences the development of the hypothalamic-adipose-axis in the embryo, especially the signaling within the arcuate and the paraventricular nuclei, influencing leptin sensitivity (Breton 2013).

18 Obesity and osteoarthritis

Treatment of overweight consists of decreasing energy intake and increasing energy expenditure. The latter is very difficult to achieve, as overweight cats are very hard to get into exercise, while obese dogs may have concurrent OA or other locomotion problems 2 (Zoran 2010). Decreasing energy intake is therefore the cornerstone of treatment of overweight in dogs and cats. Because the energy requirements of dogs and cats for maintenance, when under the same environmental conditions, varies as much as twofold (Gross et al. 2010), the weight loss plan should be based on the individual, with close monitoring. When these animals receive less food, they will slow down their metabolic rate, just like every other healthy animal; the mechanism that made them so successful during evolution (German et al. 2011, Bradshaw 2006, Speakman 2013). So owners end up with an animal that is constantly begging for food, which hardly gets any food, and still is hardly losing weight. That is the main reason why weight loss programs have a low long-term success rate (German et al. 2012). Prevention is therefore better than cure.

Osteoarthritis

In healthy joints, there is a delicate balance between anabolic and catabolic factors. Articular cartilage is compressible by its composition of collagen arcades, the extracellular matrix and interstitial water. By shock absorption during movement, the sponge-function of the articular cartilage is activated, generating diffusion for nutrients in and waste products out of the cartilage. Shock absorption takes place within the articular cartilage, as well as by bounding with the subchondral bone and periarticular structures. The synovia has a certain viscosity that lubricates the joint, delivering nutrients to the joint and transporting waste products from the joint. In osteoarthritis, there is a disturbance in the balance of anabolic and catabolic factors, in favor of catabolism, with consequences for the whole functional unit (i.e. bones, cartilage, synovia, muscles, tendons and ligaments).

Osteoarthritis (OA) is defined as a chronic degenerative joint disease, which can be divided in primary and secondary OA. Primary OA is associated with ageing. With increasing age, there is a steady increase in the synthesis of aggrecan keratin sulfate and a steady decrease in the synthesis of link proteins in cartilage, resulting in the formation of Advanced Glycation End products (AGEs, mainly pentosidine). These AGEs cause the cartilage matrix to become stiff and brittle, making it more susceptible to damage when it is mechanically loaded. Furthermore, AGEs are able to induce inflammatory pathways and stimulate the production of matrix degrading enzymes; matrix metalloproteinases (MMPs) in cartilage (Loeser et al. 2005). Secondary OA is associated with trauma, joint inflammation, or developmental orthopedic diseases. There is much debate about the existence of primary OA, as in most cases of primary OA, histologic lesions are discovered, which may be caused by microtrauma, and thus should be appointed as secondary OA (Sulzbacher 2013).

19 Part I

The imbalance in favor of catabolism is responsible for the progressive destruction of the cartilaginous extracellular matrix. When this involves the collagen arcades, this damage 2 is irreversible and has a huge impact on the shock-absorbing function of the cartilage. Collagen is degraded by biomechanical factors and MMPs, accompanied by a decreased synthesis of tissue inhibitors of MMPs (i.e. TIMPs). Many of these proteinases are released under the influence of cytokines, mainly Interleukin-1 (IL-1) and Tumor Necrosis Factor alpha (TNF-alpha). In addition, IL-1 can trigger the formation of prostaglandin E2 (PGE2). PGE2 generated in the synovial membrane induces pain, causes vasodilatation and increased vascular permeability (with more watery synovial fluid as a result). Chronic biomechanical changes at the level of the subchondral plate due to damage of the overlaying cartilage result in a reorganization of the subchondral bone, with the appearance of sclerosis and osteophyte formation at joint margins (Boire 2012), however the primary lesion being in the subchondral bone is also suggested (Sulzbacher 2013).

The damage to the cartilage matrix, the release of inflammatory mediators and proteolytic enzymes, and the decreased quality of the synovial fluid all diminish the integrity of the joint, leading to a vicious circle of cartilage degeneration (Figure 1).

Obesity and osteoarthritis

Mechanical loading is a major risk factor for OA. Obesity results in more weight bearing of the joints which predisposes an overload in the delicate balance of normal forces on cartilage and subchondral bone, finally resulting in micro-fissures. It may also cause misalignment in the major joints (hip, knee, shoulder, elbow) (Kealy et al. 1992, Kässtrom 1975, Van Hagen et al. 2005, Whitehair et al. 1993, Duval et al. 1999, Runge et al. 2008, Huck et al. 2009, Marshall et al. 2009). Although normal cyclic and intermittent weight bearing and stress on cartilage stimulates matrix production by chondrocytes, static weight baring or overstressing cartilage inhibits the production of matrix by chondrocytes (Marshall et al. 2009). In humans with preclinical knee OA, obese individuals have more severe cartilage degeneration as assessed by both morphologic and quantitative MRI measurements compared to normal weight individuals. This more severe cartilage degeneration could be caused by overloading of the joint, but the increased incidence of finger OA in obese individuals indicates that other mechanisms also play an important role (Baum et al. 2013).

It has recently become clear that adipose tissue is not only a storage form of excess calories but is also an important source of inflammatory adipokines, including leptin, IL-1, IL-6, and TNF-alpha (Fain 2010). Besides regulating appetite, leptin also has pro- inflammatory activity. Leptin is also produced in the joint, and leptin mRNA expression in cultures of chondrocytes derived from OA cartilage is higher in cartilage from obese men compared to cartilage of men of normal weight (Simopoulou et al. 2007). Interestingly, obese mice that are deficient in leptin do not develop signs of OA (Griffin et al. 2009), demonstrating the crucial role of leptin in the development of OA in overweight animals.

20 Obesity and osteoarthritis

Cartilage lesions FAT TISSUE Synovialitis 2 IL, TNFα

MMPs Watery synovial uid Cartilage degeneration WARM IL Synovia cells chondrocytes Vasodilatation Osteoclasts IL, MMP Arachidonic Acid PAINFULL

PGE2 LTB4 Inammation

Fig. 1: With primary or secondary OA as a starting point, the sequence of events depicted in this figure reveals that OA development can be blocked or slowed down by interruption of the vicious circle

10 100

) 9 8 98 7 96 rce (N/kg 6 Worst a ected forelimb

5 94 % Contralateral forelimb

ical fo 4

rt Relative body weight 3 92

ak ve 2 90

Pe 1 0 88 01428425684112 Days

Fig. 2: Weight reduction program for lame obese (i.e. 120%, of ideal weight) dogs revealed a gradual improvement of locomotion on force plate analysis of the affected leg until normal values were obtained already with a weight reduction of 6.1% (modified from Marshall et al. 2010)

21 Part I

With high leptin levels, IL-1 and TNF-alpha are produced by synoviocytes, mononuclear cells, articular cartilage, as well as by adipose tissues surrounding the joints, and can 2 up-regulate MMP-gene expression in chondrocytes (Figure 1) (Murab et al. 2013). Levels of these cytokines are elevated in joints affected by OA, and are able to induce catabolic processes in chondrocytes in vitro, leading to cartilage matrix degradation (Griffin et al. 2011). Leptin induces the production of IL-1, MMP-9, and MMP-13 in a dose-dependent manner, causing further damage to the cartilage (Simopoulou et al. 2007).

In in vivo studies with mice, in addition to changes due to OA, three additional effects of obesity with increased leptin levels were observed, namely, impaired musculoskeletal force, hyperalgesia, and mental depression. It has, therefore, been concluded that obesity may be a risk factor for inflammatory arthritis (Griffin et al. 2011, Magliano 2008) and also for additional symptoms of OA.

To date, the medical treatment of OA is limited to the prevention of PGE2 production by means of either (N)SAIDs and/or increasing the dietary intake of long-chain omega-3 fatty acids (Fernandes et al. 2002, Hielm-Björkman et al. 2012). Dietary long-chain omega-3 fatty acids that are effective in OA management are eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) EPA and DHA lower the expression of MMPs, and of pro- inflammatory cytokines like TNF-alpha and IL-1. EPA and DHA compete with arachidonic acid in the cyclooxygenase pathway in favor of the less inflammatory prostaglandins (PGs) of the 3-series and leukotrienes (LTs) of the 5-series in stead of the pro-inflammatory PGs of the 2-series and LTs of the 4-series. They should be administered in sufficient amounts (i.e. > 110.2±5.75 mg EPA+DHA per kg body weight in dogs and >146 mg EPA+DHA per kg body weight in cats) to be effective (Hielm-Björkman et al. 2012).

Another cytokine that is important in adipose tissue and is associated with OA related pain is monocyte chemoattractant protein (MCP)-1. MCP-1 is upregulated in mice after destabilization of the medial meniscus (a model for knee-OA) and acts as a ligand for the chemokine receptor 2 (CCR2). Activation of CCR2 is correlated with the presentation of movement-provoked pain behaviors, as are typically seen in OA (Miller et al. 2012). This is due to direct excitation of dorsal root ganglia neurons, as well as by attracting macrophages to the dorsal root ganglia. Current evidence confirms that patients with OA have lower pain-pressure thresholds, suggesting that central sensitization contributes to pain in OA (Suokas et al. 2012). In obese mice, as well as in obese dogs, MCP-1 is up-regulated. Feeding a high fat diet to normal weight mice, without inducing weight gain, also induces up-regulation of MCP-1, while weight loss in mice is associated with a decrease in MCP-1 (Kurki et al. 2012). This suggests positive effects of weight loss and feeding a low fat diet, on animals with OA-associated pain.

22 Obesity and osteoarthritis

OA in overweight dogs

Several studies have shown that skeletal diseases occur more often in dogs that have 2 consumed too many calories from puppyhood onward, and that the subsequent development of OA occurs more often, and is more severe, in overweight dogs. The relationship between overweight and OA has been demonstrated by Kealy et al. (2000) in Labradors, by Hedhammar et al. (1974) in Great Danes, by Kasström (1995) in German Shepherd dogs, by Brown et al. (1996) in Cocker Spaniels, and by van Hagen et al. (2005) in Boxers. This is also the experience of practicing veterinarians in many dogs of other breeds or cross-breeds. Overweight can be considered as the cause or the result of reduced activity, and in most cases it can be considered as both (Diez and Nguyen 2006).

Effects of weight loss

The effectiveness of weight reduction on OA associated pain and lameness has been reported in several studies. A study in which nine client-owned overweight dogs (>12% excess body weight) with hind leg lameness and radiographic signs of hip joint OA revealed significant improvement in body condition score and lameness, scored on a visual analogue scale after 11-18% of weight loss (Impellizeri et al. 2000). A clinical trial involving 29 overweight or obese dogs (with body condition scores of 4/5 or 5/5, respectively) with clinical evidence of one-limb lameness and radiographic signs of OA revealed a significant improvement in force plate evaluation after losing 9.3-13.6% of body weight (Mlanick et al. 2006). A prospective trial involving 14 obese (i.e. 20% overweight) client-owned dogs with lameness, demonstrated that a weight reduction program improved locomotion (measured on a visual analogue scale) when weight loss was 6.1% or more, and resulted in a better locomotion measured with force plate analysis when weight loss was 8.85% or more (Marshall et al. 2010). This demonstrates that dogs can show improvement even before they reach their optimal body weight (Figure 2).

There are several ways to achieve weight loss in dogs. Total starvation has proven to be a safe and very effective way of reducing the energy input, without negative side effects for a period up to 4 weeks in severely overweight dogs (De Bruijne 1979), with a weight reduction in obese dogs of 20% in 3 weeks (n=7 client owned dogs). However, a dual energy x-ray absorptiometry (DEXA) scan was not performed in this study to evaluate effects on body composition. Plasma albumin levels decreased, as did blood urea nitrogen, sodium, and total calcium levels. This indicates reduction of muscle mass, which is not wanted in weight loss, as a reduction in lean body mass also reduces energy expenditure post-weight loss. Currently a more gradual approach to weight loss is usually being used, because total starvation is an ethical concern. Weight reduction diets with increased protein: calorie ratio or lysine: calorie ratio enable dogs to maintain their muscle mass (Hannah and Laflamme 1998, Yamka et al. 2007). Several therapeutic diets are available that aid in achieving weight loss in dogs. The best results are achieved by using a high fiber, high protein, fat restricted diet (German et al. 2010). To optimize the

23 Part I

chance of success, the optimal body weight of the dog must be determined. This can be done by DEXA scan, bioelectrical impedance, morphometric equations, body mass index, 2 or by body condition scoring (Jeusette et al. 2012). Body condition scoring is mostly used in practice because it is easy, non-invasive and its results correlate well with DEXA analysis (Laflamme 1997). The amount of food (including treats) given to achieve gradual weight loss of 1-3% of body weight (BW) per week can be calculated by the formula: Daily Energy Requirement = Resting Energy Requirement (RER) = (optimal BW in kg)0.75 x 70 in kcal metabolizable energy per day (Linder and Freeman 2010). In cats, gradual weight loss of 0.5-2% of BW per week can be achieved when feeding 0.8xRER0.75. This is a starting point, because as food intake drops, metabolism adapts, resulting in a decreased energy requirement (German et al. 2011). Follow up is needed to prevent weight regain (German et al. 2012).

Other methods for management of OA and obesity

Another approach is the use of drugs, which are available to reduce fat absorption and should be taken with a normal food intake. Weight loss of 14.2% in overweight laboratory dogs that were administered the triglyceride protein inhibitor mitratapide without food reduction was reported without negative side effects during the study period (Dobenecker et al. 2009). We subsequently used this drug as sole therapy in overweight dogs with severe lameness, and found an improvement in locomotion, objectively registered with force plate, starting at weight reduction of 7%, without side effects (Hazewinkel et al., unpublished results).

In the future, candidate genes involved in obesity, such as melanocortin-4 receptor (MC4R) in humans, should be identified in dogs and cats, and might be developed into pharmaceutical agents (Berg, van den et al. 2010). Agonists of MC4R will lower energy intake, and increase energy expenditure (Adan et al. 2006). Weight reduction sufficient to achieve a body condition score (Laflamme 1997) of 5/9 or even lower accompanied by increased activity is most effective, but may not be desirable if the dog has OA with marked joint pain (Piermattei et al. 2006). Although owners often find it difficult to decrease the amount of food given per meal or the number of treats given between meals (Kienzle et al. 1998), they should be advised to reduce the energy intake of their dog, preferably combined with taking their dog swimming or for walks on the leash, as a way to increase energy expenditure without overloading affected joint(s). The duration of exercise should be adjusted to the effect, i.e. the dog should not become lamer after a period rest following the leash walk. NSAIDs, used to counteract the influence of pro- inflammatory mediators, not only acts as analgesics but also prevent or slow-down further joint breakdown. Omega-3 fatty acids have similar effects (Fernandes et al. 2002). Nutraceuticals, physiotechniques, intra-articular injections of anti-inflammatory drugs, periarticular gold deposits, or drugs meant to facilitate restoration of damaged cartilage might have a place in some cases, however, up till now there is not enough evidence to support the use of these products (Vandeweerd et al. 2012). In cases of severe OA,

24 Obesity and osteoarthritis

when weight reduction, daily dosage of NSAIDs, and life style adaptations do not lead to sufficient improvement, invasive surgery may be an option. 2 Conclusion

Obesity can play an important role in OA development. For OA prevention, maintaining normal weight is important, because weight gain promotes OA development. For treatment, weight loss is the most important aspect of achieving a long-lasting improvement in locomotion in dogs that are lame as a result of OA.

25 Part I

References • Breton, C., 2013. The hypothalamus-adipose axis is a key target of developmental 2 • Adan, R.A.H., Tiesjema, B., Hillebrand, J.J.G., programming by maternal nutritional Fleur, S.E. la, Kas, M.J.H., Krom, M. de, 2006. manipulation. Journal of Endocrinology 216, The MC4 receptor and control of appetite. R19-R32. British Journal of Pharmacology 149, 815–827. • Brown, D.C., Conzemius, M.G., Shofer, F.S., • Alexander, L.G., Salt, C., Thomas, G., 1996. Body weight as a predisposing factor for Butterwick, R., 2011. Effects of neutering humeral condylar fractures, cranial cruciate on food intake, body weight and body rupture and intervertebral disc disease in composition in growing female kittens. Cocker Spanials. Veterinary and Comparative The British Journal of Nutrition 106 Suppl 1, Orthopaedics and Traumatology 9, 75-78. S19-23. • Brown, M., 2008. Skeletal muscle and bone: • Baum, T., Joseph, G.B., Nardo, L., Virayavanich, effect of sex steroids and aging. Advances in W., Arulanandan, A., Alizai, H., Carballido- Physiology Education 32, 120–126. Gamio, J., Nevitt, M.C., Lynch, J., McCulloch, • Bruijne, J.J. de, 1979. Biochemical C.E., Link, T.M., 2013. Correlation of magnetic observations during total starvation in dogs. resonance imaging–based knee cartilage T2 International Journal of Obesity 3, 239-247. measurements and focal knee lesions with • Cave, N.J., Backus, R.C., Marks, S.L., Klasing, body mass index: Thirty-six–month followup K.C., 2007. Oestradiol, but not genistein, data from a longitudinal, observational inhibits the rise in food intake following multicenter study. Arthritis Care & Research gonadectomy in cats, but genistein is 1, 23–33. associated with an increase in lean body • Becker, N., Dillitzer, N., Sauter-Louis, C., mass. Journal of Animal Physiology and Kienzle, E., 2012. Feeding of dogs and cats in Animal Nutrition 91, 400-410. Germany. Tierarztliche Praxis 40, 391-397. • Colliard, L., Paragon, B.M., Lemuet, B., Bénet, • Begg, D.P., Woods, S.C., 2013. The J.-J., Blanchard, G., 2009. Prevalence and risk endocrinology of food intake. Nature Reviews factors of obesity in an urban population of Endocrinology 9, 584-597. healthy cats. Journal of Feline Medicine and • Berg, L. van den, Berg, S.M. van den, Martens, Surgery 11, 135-140. E.E.C.P., Hazewinkel, H.A.W., Dijkshoorn, • Courcier, E.A., Mellor, D.J., Thomson, R.M., N.A., Delemarre-van de Waal, H.A., Heutink, Yam, P.S., 2011. A cross sectional study of P., Leegwater, P.A.J., Heuven, H.C.M., 2010. the prevalence and risk factors for owner Analysis of variation in the melanocortin-4 misperception of canine body shape in first receptor gene (mc4r) in golden retriever dogs. opinion practice in Glasgow. Preventive Animal genetics, 41, 557-560. Veterinary Medicine 102, 66-74. • Boire, C., 2012. Update on the • Diez, M., Nguyen, P.,2006. Obesity: pathophysiology of osteoarthritis in dogs. epidemiology, pathophysiology and Point Veterinaire 43, 62-68. management of the obese dog in: • Bradshaw, J.W.S., 2006. The evolutionary basis Encyclopedia of canine clinical Nutrition, for the feeding behavior of domestic dogs Ed by Pibot P, Biourge V, Elliot D. Aniwa SAS, (Canis familiaris) and cats (Felis catus). Journal France, 2-51. of Nutrition 136, 1927S-1931S.

26 Obesity and osteoarthritis

• Dobenecker, B., Bock, M. de, Engelen, M., • German, A.J., Holden, S.L., Morris, P.J., Biourge, Goossens, L., Scholz, A., Kienzle, E., 2009. V., 2012. Long-term follow-up after weight Effect of mitratapide on body composition, management in obese dogs: The role of diet 2 body measurements and glusose tolerance in preventing regain. The Veterinary Journal in obese Beagles. Veterinary Research 192, 65-70. Communications 33, 839-847. • Grant, R.W., Vester Boler, B.M., Ridge, T.K., • Duval, J.M., Budsberg, S.C., Flo, G.L., Graves, T.K., Swanson, K.S. Adipose tissue Sammarco, J.L., 1999. Breed, sex, and body transcriptome changes during obesity weight as risk factors for rupture of the cranial development in female dogs. Physiological cruciate ligament in young dogs. Journal of Genomics 43, 295-307. the American Veterinary Medical Association • Griffin, T.M., Fermor, B., Huebner, J.L., Kraus, 215, 811-814. V.B., Rodriguis, R.M., Wetsel, W.C., Cao, L., • Fain, J.N., 2010. Release of inflammatory Setton, L.A., Guilak, F., 2011. Diet-induced mediators by human adipose tissue is obesity differentially regulates behavioural, enhanced in obesity and primarily by the biomechanical, and molecular risk factors for nonfat cells: a review. Hindawi Publishing OA in mice. Arthritis Reseach Therapy 12, R13. Corporation, ID 513948. • Griffin, T.M., Huebner, J.L., Kraus, V.B., Guilak, • Feng, H., Zheng, L., Feng, Z., Zhao, Y., Zhang, F., 2009. Extreme obesity due to impaired N., 2013. The role of leptin in obesity and the leptin signalling in mice does not cause knee potential for leptin replacement therapy. osteoarthritis. Arthitis & Rheumatology 60, Endocrine 44, 33-39. 2935-2944. • Fernandes, J.C., Martel-Pelletier, J., • Gross, K.L., Yamka, R.M., Khoo, C., Friesen, K.G., Pelletier, J.P., 2002. The role of cytokines in Jewell, D.E., Schoenherr, W.D., Debraeckeleer, osteoarthritis pathophysiology. Biorheology J., Zicker, S.C., 2010. Macronutrients. In: Small 39 237-246. Animal Clinical Nutrition 5th Ed. Hand et al. • Geary, N., Asarian, L., Korach, K.S., Pfaff, D.W., Eds, Mark Morris Institute, Topeka, USA, p. 62. Ogawa, S., 2001. Deﬁcits in E2-dependent • Hagen, M.A. van, Ducro, B.J., Broek, J. van control of feeding, weight gain, and den, Knol, B.W., 2005. Incidence, risk factors, cholecystokinin satiation in ER-alpha null and hereditability estimates of hind limb mice. Endocrinology 142, 4751–4757. lameness caused by hip dysplasia in a birth • German, A.J., 2006. The growing problem of control cohort of Boxers. American Journal of obesity in dogs and cats. Journal of Nutrition Veterinary Research 66, 307-312. 136, 1940S-1946S. • Hannah, S.S., Laflamme, D.P., 1998. Increased • German, A.J., Holden, S.L., Bissot, T., Morris, dietary protein spares lean body mass during P.J., Biourge, V., 2010. A high protein high fibre weight loss in dogs. Journal of Veterinary diet improves weight loss in obese dogs. The Internal Medicine 12, 224-230. Veterinary Journal 183, 294-297. • Hedhammar, A., Fu-Ming, W., Krook, L., • German, A.J., Holden, S.L., Mather, N.J., Morris, Schryver, H.F., De Lahunta, A., Whalen, J.P., P.J., Biourge, V., 2011. Low-maintenance Kallfelz, F.A., Nunez, E.A., Hintz, H.F., Sheffy, energy requirements of obese dogs after B.E., Ryan, G.D., 1974. Overnutrition and weight loss. The British journal of nutrition skeletal disease; an experimental study 106(Suppl 1), S93-96. in growing Great Dane dogs. The Cornell Veterinarian 64 (Suppl. 5), 1- 135.

27 Part I

• Hielm-Björkman, A., Roine, J., Elo, K., • Kasström, H., 1975. Nutrition, weight gain Lappalainen, A., Junnila, J., Laitinen- and development of hip dysplasia. An 2 Vapaavuori, O., 2012. An un-commissioned experimental investigation in growing dogs randomized, placebo-controlled double- with special reference to the effect of feeding blind study to test the effect of deep sea fish intensity. Acta Radiologica Supplementum oil as a pain reliever for dogs suffering from 344, 135-179. canine OA. BMC Veterinary Research 8, Article • Kealy, R.D., Lawler, D.F., Ballam, J.M., Lust, number 157. G., Biery, D.N., Smith, G.K., Mantz, S.L., 2000. • Houpt, K.A., 1991. Feeding and drinking Evaluation of the effect of limited food behavior problems. Veterinary Clinics of North consumption on radiographic evidence America - Small Animal Practice 21, 281-298. of osteoarthritis in dogs. Journal of the • Huck, J.L., Biery, D.N., Lawler, D.F., Gregor, T.P., American Veterinary Medical Association Runge, J.J., Evand, R.H., Kealy, R.D., Smith. G.K., 217, 1678-1680. 2009. A longitudinal study of the influence of • Kealy, R.D., Lawler, D.F., Ballum, J.M., Mantz, lifetime food restriction on development of S.L., Biery, D.N., Greely, E.H., Lust, G., Serge, osteoarthritis in the canine elbow. Veterinary M., Smith, G.K., Stowe, H.D., 2002. Effect of Surgery 38, 192-198. diet restriction on life span and age-related • Impellizeri, J.A., Tetrick, M.A., Muir, P., 2000. changes in dogs. Journal of the American Effect of weight reduction on clinical signs Veterinary Medical Association 220, of lameness in dogs with hip osteoarthritis. 1315-1320. Journal of the American Veterinary Medical • Kealy, R.D., Olsson, S.E., Monti, K.L., Lawler, Association 216, 1089-1091. D.F., Biery, D.N., Helms, R.W., Lust, G., Smith, • Ishioka, K., Soliman, M.M., Sagawa, M., G.K., 1992. Effects of limited consumption Nakadomo, F., Shibata, H., Honjoh, T., on the incidence of hip dysplasia in growing Hashimoto, A., Kitamura, H., Kimura, K., Saito, dogs. Journal of the American Veterinary M., 2002. Experimental and clinical studies Medical Association 201, 857-863. on plasma leptin in obese dogs. Journal of • Kienzle, E., Bergler, R., Mandernach, A., 1998. Veterinary Medical Science 64, 349-353. Comparison of the feeding behaviour of the • Jeusette, I., Detilleux, J., Cuvelier, C., Istasse, L., man-animal relationship in owners of normal Diez, M., 2004. Ad libitum feeding following and obese dogs. Journal of Nutrition 128, ovariectomy in female Beagle dogs: Effect 2777S-2782S. on maintenance energy requirement and • Kurki, E., Shi, J., Martonen, E., Finckenberg, P., on blood metabolites. Journal of Animal Mervaala, E., 2012. Distinct effects of calorie Physiology and Animal Nutrition 88, 117-121. restriction on adipose tissue cytokine and • Jeusette, I., Greco, D., Aquino, F., Detilleux, angiogenesis profiles in obese and lean J., Peterson, M., Romano, V., Torre, C., 2012. mice. Nutrition and Metabolism 2012, Article Effect of breed on body composition and number 64. comparison between various methods to • Laflamme, D.P., 1997. Development and estimate body composition in dogs. Research validation of a body condition score system in Veterinary Science 88, 227-232. for dogs. Canine Practice 22, 10-15.

28 Obesity and osteoarthritis

• Laflamme, D.P., 2005. Nutrition for Aging • Meli, R., Pacilio, M., Raso, G.M., Esposito, E., Cats and Dogs and the Importance of Coppola, A., Nasti, A., Di Carlo, C., Nappi, C., Body Condition. Veterinary Clinics of North Di Carlo, R., 2004. Estrogen and Raloxifene 2 America: Small Animal Practice 35, 713-742. modulate leptin and its High-protein diet • Linder, D.E., Freeman, L.M., 2010. Evaluation following ovariohysterectomy receptor of calorie density and feeding directions for in hypothalamus and adipose tissue from commercially available diets designed for ovariectomized rats. Endocrinology 145, weight loss in dogs and cat. Journal of the 3115–3121. American Veterinary Medical Association • Mlacnik, E., Bockstahler, B.A., Müller, M., 236, 74-77. Tetrick, M.A., Nap, R.C., Zentek, J., 2006. Effects • Loeser, R.F., Yammani, R.R., Carlson, C.S., Chen, of caloric restriction and a moderate or H., Cole, A., Im, H.J., Bursch, L.S., Yan, S.D., intense physiotherapy program for treatment 2005. Articular chondrocytes express the of lameness in overweight receptor for advanced glycation end products dogs with osteoarthritis. Journal of the : potential role in osteoarthritis. Arthritis & American Veterinary Medical Association Rheumatology 52, 2376-2385. 229, 1756-1760. • Magliano M., 2008. Obesity and arthritis. • Murab, S., Chameettachal, S., Bhattacharjee, Menopause International 14, 149-154. M., Das, S., Kaplan, D.L., Ghosh, S., • Marshall, W.G., Bockstahler, B., Hulse, D. 2013. Matrix-embedded cytokines to Carmichael, S.R., 2009. Review of osteoarhritis simulate osteoarthritis-like cartilage and obesity: current understanding of the microenvironments. Tissue Engineering - relationship and benefit of obesity treatment Part A 19, 1733-1753. and prevention in the dog. Veterinary and • Nijland, M.L., Stam, F., Seidell, J.C., 2010. Comparative Orthopaedics and Traumatology Overweight in dogs, but not in cats, is related 5, 339-345. to overweight in their owners. Public Health • Marshall, W.G., Hazewinkel, H.A., Mullen, D., Nutrition 13, 102-106. Meyer, G. de, Baert, K., Carmichael, S., 2010. • Oda-Montecinos, C., Saldaña, C., Andrés, A., The effect of weight loss on lameness in 2013. Eating behaviors are risk factors for obese dogs with osteoarthritis. Veterinary the development of overweight. Nutrition Research Communications 34, 241-253. Research 33, 796-802. • McGreevy, P.D., Thomson, P.C., Pride, C., • Piermattei, D.L., Flo, G.L., DeCamp, C.E., 2006. Fawcett, A., Grassi, T., Jones, B., 2005. Small animal orthopaedics and fracture repair, Prevalence of obesity in dogs examined by Ch. 16: The hip joint, Saunders, St. Louis, Australian veterinary practices and the risk (USA), pp. 433- 511. factors involved. Veterinary Record 156, • Racette, S.B., Deusinger, S.S., Deusinger, 695-702. R.H., 2003. Obesity: Overview of prevalence, etiology, and treatment. Physical Therapy 83, 276-288. • Ravelli, A.C., Meulen, J.H. van der, Osmond, C., Barker, D.J., Bleker, O.P., 1999. Obesity at the age of 50 y in men and women exposed to famineprenatally. American Journal of Clinical Nutrition 70, 811–816.

29 Part I

• Robertson, I.D., 2003. The association of • Speakman, J.R., 2013. Evolutionary exercise, diet and other factors owner- perspectives on the obesity epidemic: 2 perceived obesity in privately owned dogs Adaptive, maladaptive, and neutral from metropolitan Perth WA. Preventive viewpoints. Annual Review of Nutrition 33, Veterinary Medicine 58, 75-83. 289-317. • Rosenbaum, M., Leibel, R.L., 1999. Clinical • Sulzbacher, I., 2013. Osteoarthritis: Histology review 107: Role of gonadal steroids in the and pathogenesis. Wiener Medizinische sexual dimorphisms in body composition and Wochenschrift 163, 212-219. circulating concentrations of leptin. Journal • Suokas, A.K., Walsh, D.A., McWilliams, D.F., of Clinical Endocrinology and Metabolism 84, Condon, L., Moreton, B., Wylde, V., Arendt- 1784-1789. Nielsen, L., Zhang, W., 2012. Quantitative • Rothwell, N.J., Stock, M.J., 1979. Role for sensory testing in painful osteoarthritis: brown adipose tissue in diet-induced A systematic review and meta-analysis. thermogenesis. Nature 281, 31–35. Osteoarthritis Cartilage 20, 1075-1085. • Runge, J.J., Bierry, D.N., Lawler, D.F., Gregor, • Vandeweerd, J.-M., Coisnon, C., Clegg, T.P., Evans, R.H., Kealy, R.D., Szabo, S.D., P., Cambier, C., Pierson, A., Hontoir, F., Smith, G.K., 2008. The effects of lifetime Saegerman, C., Gustin, P., Buczinski, S., 2012. food restriction on the development of Systematic review of efficacy of nutraceuticals osteoarthritis in the canine shoulder. to alleviate clinical signs of osteoarthritis. Veterinary Surgery 32, 102-107. Journal of Veterinary Internal Medicine 26, • Salomon, J.A., Wang, H., Freeman, M.K., Vos, T., 448-456. Flaxman, A.D., Lopez, A.D., Murray, C.J., 2013. • Vis, B., Hylands, T., 2013. Fat government, thin Healthy life expectancy for 187 countries, populace? Is the growth of obesity prevalence 1990-2010: a systematic analysis for the lower in more generous welfare states? Global Burden Disease Study 2010. Lancet International Journal of Social Welfare 22, 380, 2144-2162. 360-373. • Simopoulou, T., Malizos, K.N., Iliopoulos, D., • Wang, Y., Wluka, A.E., Simpson, J.A., Giles, Stefanou, N., Papatheodorou, L., Ioannou, M., G.G., Graves, S.E., de Steiger, R.N., Cicuttini, Tsezou, A., 2007. Differential expression of F.M., 2013. Body weight at early and middle leptin and leptin’s receptor isoform (Ob-Rb) adulthood, weight gain and persistent mRNA between advanced and minimally overweight from early adulthood are affected OA cartilage; effect on cartilage predictors of the risk of total knee and hip metabolism. Osteoarthritis Cartilage 15, replacement for osteoarthritis. Rheumatology 872-883. 52, 1033-1041. • Slingerland, L.I., Hazewinkel, H.A., Meij, B.P., • Whitehair, J.G., Vasseur, P.B., Willits, N.H., 1993. Picavet, P., Voorhout, G., 2011. Cross-sectional Epidemiology of cranial cruciate ligament study of the prevalence and clinical features rupture in dogs. Journal of the American of osteoarthritis in 100 cats. The Veterinary Veterinary Medical Association 203, Journal 187, 304-309. 1016-1019.

30 Obesity and osteoarthritis

• Yamka, R.M., McLeod, K.R., Harmon, D.L., Freetly, H.C., Schoenherr, W.D., 2007. The impact of dietary protein source on observed 2 and predicted metabolizable energy of dry extruded dog foods. Journal of Animal Science 85, 204-212. • Zoran, D.L., 2010. Obesity in Dogs and Cats: A Metabolic and Endocrine Disorder. Veterinary Clinics of North America - Small Animal Practice 40, 221-239.

31 32 Chapter 3

Obesity in show dogs

Ronald J. Corbee

Journal of Animal Physiology and Animal Nutrition 2013; Volume 97, Issue 5, Pages 904-910

33 Part I

Abstract

Obesity is an important disease with a growing incidence. Because obesity is related to several other diseases, and decreases life 3 span, it is important to identify the population at risk. Several risk factors for obesity have been described in the literature. A higher incidence of obesity in certain breeds is often suggested. The aim of this study was to determine whether obesity occurs more often in certain breeds. The second aim was to relate the increased prevalence of obesity in certain breeds to the official standards of that breed. To this end we investigated 1379 dogs of 128 different breeds by determining their body condition score (BCS). Overall, 18.6% of the show dogs had a BCS >5, and 1.1% of the show dogs had a BCS>7. There were significant differences between breeds, which could be correlated to the breed standards. It warrants firm discussions with breeders and judges in order to come to different interpretations of the standards in order to prevent overweight conditions from being the standard of beauty.

34 Obesity in show dogs

Introduction

Obesity is an important disease with a growing incidence from 33-35% in the 90s (Armstrong and Lund 1996) up to 40-60% in more recent studies (Courcier et al. 2010, McGreevy et al. 2005, Ricci et al. 2007). Because obesity is related to several other diseases 3 (i.e. osteoarthritis, cardiovascular disease and diabetes mellitus; Laflamme 2005) and decreases life span (Kealy et al. 2002) it is important to identify the population at risk, in order to take measures to prevent occurrence of the disease. Several risk factors for obesity have been described in the literature. Increasing age is associated with decreased activity (Laflamme 2005) and a decrease in lean body mass in dogs (Kealy et al. 1999). Neutered dogs are more likely to become obese by a decrease in maintenance energy requirements and an increase in spontaneous food intake (Laflamme 2005). Other risk factors for obesity in dogs are type of diet, socio-economic status of the owner, changes in lifestyle, and body condition of the owner (Colliard et al. 2006, Courcier et al. 2010, Nijland et al. 2010).

Several authors mention breed as a risk factor and suggest a higher incidence of obesity in certain breeds (Ishioka et al. 2007, Jeusette et al. 2010, McGreevy et al. 2005); however they could not demonstrate sufficient data to confirm their observations because of the small number (n=19) of dogs (Jeusette et al. 2010), a small number (n=5) of breeds investigated (Ishioka et al. 2007), or designation to nine non-standardized breed groups (McGreevy et al. 2005). Significant differences were demonstrated between breed groups, and crossbred dogs were most likely to become obese in one study (McGreevy et al. 2005). Breed differences were demonstrated in body composition by using different methods (i.e. Dual-Energy X-Ray Absorptiometry, Bioelectrical Impedance, morphometric equations, body condition score and body mass index), which suggests dogs of certain breeds have an increased risk of becoming obese (Jeusette et al. 2010). Significantly higher risk for overweight conditions in the Labrador Retriever (n=25) was demonstrated in a survey of 616 healthy dogs visiting veterinary clinics (results for other breeds were not specified) (Colliard et al. 2006).

Breed standards have an impact on the visual appearance of a dog, and may also influence body condition. Our hypothesis is that the body condition score (BCS) of breeds that are to be very robust and muscular, as defined by the breed standards (i.e. Bernese Mountain dog or Mastiff), is higher, whereas other breeds are expected to have a lower BCS because they are to be lean and elegant, as defined by the breed standards (i.e. Greyhound and Whippet). The aim of this study is to determine whether overweight conditions occur more often in certain breeds. The second aim is to relate the increased prevalence of overweight conditions in certain breeds to the official standards of that breed.

35 Part I

Materials and methods

At the Winner dog show, 1379 dogs (from 128 breeds) were investigated by determining their BCS on a 9-point scale (Laflamme et al. 1997) by inspection and palpation. All dogs 3 were scored by a board certified nutritionist (R.C.). Not all the dogs that visited the Winner dog show were scored, because some owners did not give their permission to palpate their dogs. Despite the exclusion of some dogs, the large number of dogs included is still regarded as representative for the total population present at the Winner dog show. Body condition score was used to be able to score, uniformly, a large number of dogs of a variety of breeds. The effects of breed and breed groups on BCS were investigated. Breed groups were defined by the Fédération Cynologique International (FCI). Breed standards were obtained from the American Kennel Club and were investigated for descriptions that may predispose dogs to become obese.

Statistical analysis A Kolmogorov-Smirnov test was performed to test for normality. Because the BCS in most groups was not normally distributed, a non-parametric test was used. A Kruskal-Wallis test was performed to demonstrate differences between breeds. The scores of the breeds were tested against the median value of all show dogs. Differences between sexes were tested within breeds using a Mann-Whitney-U test. A p-value of 0.05 was set as the level of significance. A p-value of <0.10 was defined as a trend. Overweight was defined as a BCS >5 on a 9-point scale. Obesity was defined as a BCS >7 on a 9-point scale. Based on a power-analysis a breed group should consist of at least 5 dogs in order to demonstrate significant differences with an alpha of 0.05 and a beta of 0.10. Breeds and/or breed groups with lower numbers of dogs were excluded from the results, but were included in the overall calculation of the prevalence of overweight conditions in show dogs.

Results BCS The dogs visiting dog shows demonstrated an incidence of overweight (i.e. a BCS>5 on a 9-point scale) of 18.6%, and an incidence of obesity (i.e. a BCS>7 on a 9-point scale) of 1.1%. The average BCS of all investigated show dogs (n=1379) was 4.67±1.69 on a 9-point scale. There were significant differences between breeds (Table 1). Some breed groups had a significantly higher BCS (i.e. Molossoid breeds, Swiss Mountain and Cattle dogs, Asian spitz and related breeds, Scenthounds, Retrievers, Water dogs, Bichons and related breeds) whereas some breed groups had a significantly lower BCS (i.e. Sighthounds) (Table 2). There were no significant differences between sexes.

Breed standards Breed standards (i.e. phrases from the general appearance) of the breeds with significantly higher (p<0.05) body condition score

36 Obesity in show dogs

Table 1: Mean Body Condition Score (BCS) of show dogs of different breeds

Breed Number of dogs Mean BCS + range Great Dane 13 3.92 (3-5) * Greyhound 5 4.00 (4) * 3 Italian Greyhound 5 4.00 (4) * Whippet 28 4.08 (3-5) * Borzoi 30 4.17 (3-5) * Boxer 14 4.20 (3-5) * Irish Setter 29 4.21 (4-5) * Dobermann 18 4.28 (3-5) * Irish Wolfshound 11 4.36 (4-6) * Rhodesian Ridgeback 31 4.39 (4-6) * English Setter 10 4.40 (4-6) * Cairn terrier 18 4.44 (4-5) Weimaraner 21 4.48 (4-6) * Airdale Terrier 7 4.50 (4-5) Pyrenian Shepherd 6 4.50 (4-5) Viszla 9 4.55 (4-5) Afghan Hound 27 4.55 (3-7) Saarloos Wolfhound 21 4.57 (4-6) Akita Inu 5 4.60 (4-6) American Cocker Spaniel 5 4.60 (4-6) Pyrenean Mountain Dog 8 4.63 (4-5) Bordeaux Dog 6 4.67 (4-5) Cavelier King Charles Spaniel 18 4.67 (3-6) Miniature Schnauzer 18 4.67 (4-6) Average 1379 4.67 (2-8) Gordon Setter 38 4.67 (4-7) Shar-Pei 12 4.67 (4-5) English Cocker Spaniel 18 4.72 (4-6) Flatcoated Retriever 23 4.73 (4-6) Markiesje 8 4.75 (4-6) German Pointing Dog 45 4.76 (4-7) German Miniature Pinscher 6 4.83 (4-5) Bearded Collie 12 4.83 (4-6) English Springer Spaniel 14 4.86 (4-7) Scottish Shepherd 5 4.88 (4-7)

37 Part I

Table 1: Continued

Breed Number of dogs Mean BCS + range Norwich Terrier 9 4.89 (4-6) 3 Rottweiler 9 4.91 (4-6) Dalmatian 23 4.91 (4-7) Australian Shepherd 38 4.96 (4-6) Boston terrier 8 5.00 (4-6) Appenzell Mountain Dog 11 5.00 (4-6) Golden Retriever 38 5.00 (4-7) T Kooikerhondje 12 5.00 (4-6) T Mastiff 10 5.10 (4-6) * Bouvier des Flandres 21 5.10 (4-6) T Keeshond 6 5.17 (4-6) * Basset 23 5.17 (4-7) * Miniature Dachshund 15 5.20 (4-6) * Labrador Retriever 45 5.24 (4-8) * Nova Scotia Duck Tolling Retriever 20 5.25 (4-7) * Bull terrier 13 5.30 (4-6) * Bullmastiff 15 5.33 (4-7) * Wetterhoun 8 5.38 (5-6) * Leonberger 11 5.44 (4-7) * Beagle 11 5.45 (4-6) * Sussex Spaniel 7 5.57 (5-7) * English Bulldog 10 5.60 (4-7) * Giant Schnauzer 10 5.60 (5-7) * Kelpie 8 5.62 (4-7) * Chow chow 8 5.75 (5-7) * Bernese Mountain Dog 14 5.79 (4-8) * Pug 6 6.00 (6) * Newfoundland 12 6.17 (5-7) *

* = p<0.05 Kruskal-Wallis test T = p<0.10 Kruskal-Wallis test

38 Obesity in show dogs

Table 2: Mean Body Condition Score BCS of show dogs of different FCI groups

FCI group section Number of dogs Mean BCS + range 2.3 Swiss Mountain dogs 25 5.44 (4-8) * 9.6 Chihuahua's 11 5.36 (4-7) * 3 6.1 Scenthounds 46 5.13 (4-7) * 9.11 Small Molossion Type dogs 20 5.10 (4-7) * 8.3 Water dogs 18 5.06 (4-8) * 8.1 Retrievers 137 5.04 (4-8) * 2.2 Molossoid breeds 111 5.02 (4-8) * 5.5. Asian Spitz and related breeds 18 4.94 (3-7) * 5.4 European Spitz 11 4.91 (4-6) 1.1 Sheepdogs 276 4.89 (4-7) 3.3 Bull-type terriers 36 4.89 (4-7) 8.2 Flushing dogs 75 4.88 (4-7) 1.2 Cattle dogs 29 4.76 (4-8) 9.1 Bichon Frisé 8 4.75 (4-7) 2.1 Pinschers and Schnauzers 91 4.67 (4-7) 7.1 Continental Pointing dogs 130 4.67 (4-7) 9.7 English Toy Spaniels 18 4.67 (4-6) 3.2 Small-sized terriers 28 4.61 (4-7) 6.3 Dalmatian and Rhodesian Ridgeback 54 4.61 (3-7) 4.0 Dachshunds 27 4.59 (4-7) 3.1 Large- and medium-sized terriers 22 4.55 (4-7) 7.2 British and Irish Pointers and Setters 67 4.43 (3-7) 10.1 Longhaired Sighthounds 59 4.37 (3-7) * 10.2 Roughhaired Sighthounds 11 4.36 (3-6) * 9.2 Poodles 10 4.30 (3-6) * 10.3 Shorthaired Sighthounds 30 4.10 (3-5) *

* p<0.05 Kruskal-Wallis test

39 Part I

Basset It is a short-legged dog, heavier in bone, size considered, than any other breed of dog, and while its movement is deliberate, it is in no sense clumsy. The average (range) BCS was 5.17 (4-7) in 23 dogs. 3 Beagle A miniature Foxhound, solid and big for his inches, with the wear-and-tear look of the hound that can last in the chase and follow his quarry to the death. The average (range) BCS was 5.45 (4-6) in 11 dogs.

Bull Terrier The Bull Terrier must be strongly built, muscular, symmetrical and active, with a keen determined and intelligent expression, full of fire but of sweet disposition and amenable to discipline. The average (range) BCS was 5.30 (4-6) in 13 dogs.

Bullmastiff That of a symmetrical animal, showing great strength, endurance, and alertness; powerfully built but active. The average (range) BCS was 5.33 (4-7) in 15 dogs.

Chow chow A powerful, sturdy, squarely built, upstanding dog of Arctic type, medium in size with strong muscular development and heavy bone. The average (range) BCS was 5.75 (5-7) in 8 dogs.

English bulldog The perfect Bulldog must be of medium size and smooth coat; with heavy, thick-set, low- swung body, massive short-faced head, wide shoulders and sturdy limbs. The average (range) BCS was 5.60 (4-7) in 10 dogs.

Giant Schnauzer Robust, strongly built, nearly square in proportion of body length to height at withers, active, sturdy, and well muscled. Temperament which combines spirit and alertness with intelligence and reliability. The average (range) BCS was 5.60 (5-7) in 10 dogs.

Keeshond The general appearance shall be that of a lithe, active dog of great quality, showing hard muscular condition combined with great suppleness of limb and conveying the capability of untiring work. It must be free from any suggestion of weediness. The average (range) BCS was 5.17 (4-6) in 6 dogs.

40 Obesity in show dogs

Labrador Retriever The Labrador Retriever is a strongly built, medium-sized, short-coupled, dog possessing a sound, athletic, well-balanced conformation that enables it to function as a retrieving gun dog. The average (range) BCS was 5.24 (4-8) in 45 dogs. Obesity (BCS of 8 on a 9-point scale) 3 was present in 6 dogs, from which one was awarded by the judge in the top-4 in its class.

Leonberger He is distinguished by his balanced build, black mask, and double coat. Adult males, in particular, are powerful and strong and carry a lion–like mane on the neck and chest. A dog or bitch is easily discernable as such. The average (range) BCS was 5.44 (4-7) in 11 dogs.

Mastiff The Mastiff is a large, massive, symmetrical dog with a well-knit frame. The impression is one of grandeur and dignity. Dogs are more massive throughout. The average (range) BCS was 5.10 (4-6) in 10 dogs.

Miniature Dachshund Low to ground, long in body and short of leg, with robust muscular development; the skin is elastic and pliable without excessive wrinkling. The average (range) BCS was 5.20 (4-6) in 15 dogs.

Newfoundland The Newfoundland is a large, heavily coated, well balanced dog that is deep-bodied, heavily boned, muscular, and strong. The average (range) BCS was 6.17 (5-7) in 12 dogs.

Nova Scotia Duck Tolling Retriever This medium sized, powerful, compact, balanced dog is the smallest of the retrievers. The average (range) BCS was 5.25 (4-7) in 20 dogs.

Pug Symmetry and general appearance are decidedly square and cobby. A lean, leggy Pug and a dog with short legs and a long body are equally objectionable. The average (range) BCS was 6.00 (6) in 6 dogs.

Sussex spaniel Its short legs, massive build, long body, and habit of giving tongue when on scent made the breed ideally suited to penetrating the dense undergrowth and flushing game within range of the gun. The average (range) BCS was 5.57 (5-7) in 7 dogs.

41 Part I

Wetterhoun A well balanced dog, traditionally used for otter hunting. A sturdy animal, but neither plump nor clumsy, square and thick set in overall build, with close fitting skin, free from throatiness or dewlap. 3 The average (range) BCS was 5.38 (5-6) in 8 dogs.

Bernese Mountain Dog He is sturdy and balanced. Dogs appear masculine, while bitches are distinctly feminine. The average (range) BCS was 5.79 (4-8) in 14 dogs. Obesity (BCS of 8 on a 9-point scale) was present in 3 dogs.

Breed standards of breeds with a tendency (p<0.10) to a higher body condition score

Bouvier des Flandres (p=0.074) The Bouvier des Flandres is a powerfully built, compact, short-coupled, rough-coated dog of notably rugged appearance. The average (range) BCS was 5.10 (4-6) in 21 dogs.

Golden retriever (p=0.054) A symmetrical, powerful, active dog, sound and well put together, not clumsy nor long in the leg, displaying a kindly expression and possessing a personality that is eager, alert and self-confident. The average (range) BCS was 5.00 (4-7) in 38 dogs.

Kooikerhondje (Dutch Decoy Spaniel) (p=0.088) The Kooikerhondje is a harmoniously built orange-red parti-coloured small sporting dog of almost square body proportions. The average (range) BCS was 5.00 (4-6) in 12 dogs.

Breed standards of the breeds with significantly lower body condition score

Borzoi The Borzoi should always possess unmistakable elegance, with flowing lines, graceful in motion or repose. Males, masculine without coarseness; bitches, feminine and refined. The average (range) BCS was 4.17 (3-5) in 30 dogs.

Boxer The ideal Boxer is a medium-sized, square-built dog of good substance with short back, strong limbs, and short, tight-fitting coat. His well-developed muscles are clean, hard, and appear smooth under taut skin. His movements denote energy. The average (range) BCS was 4.20 (3-5) in 14 dogs.

42 Obesity in show dogs

Dobermann Pinscher The appearance is that of a dog of medium size, with a body that is square. Compactly built, muscular and powerful, for great endurance and speed. Elegant in appearance. The average (range) BCS was 4.28 (3-5) in 18 dogs. 3 English Setter An elegant, substantial and symmetrical gun dog suggesting the ideal blend of strength, stamina, grace, and style. The average (range) BCS was 4.40 (4-6) in 10 dogs.

Great Dane The Great Dane combines, in its regal appearance, dignity, strength and elegance with great size and a powerful, well-formed, smoothly muscled body. The average (range) BCS was 3.92 (3-5) in 13 dogs.

Greyhound Strongly built, upstanding, of generous proportions, muscular power and symmetrical formation, with long head and neck, clean well laid shoulders, deep chest, capacious body, slightly arched loin, powerful quarters, sound legs and feet, and a suppleness of limb, which emphasize in a marked degree its distinctive type and quality. The average (range) BCS was 4.00 (4) in 5 dogs.

Irish Setter The Irish Setter is an active, aristocratic bird dog, rich red in color, substantial yet elegant in build. The average (range) BCS was 4.21 (4-5) in 29 dogs.

Irish Wolfhound Of great size and commanding appearance, the Irish Wolfhound is remarkable in combining power and swiftness with keen sight. The largest and tallest of the galloping hounds, in general type he is a rough-coated, Greyhound-like breed; very muscular, strong though gracefully built. The average (range) BCS was 4.36 (4-6) in 11 dogs.

Italian Greyhound The Italian Greyhound is very similar to the Greyhound, but much smaller and more slender in all proportions and of ideal elegance and grace. The average (range) BCS was 4.00 (4) in 5 dogs.

Rhodesian ridgeback The Ridgeback represents a strong, muscular and active hound, symmetrical and balanced in outline. A mature Ridgeback is a handsome, upstanding and athletic dog. The average (range) BCS was 4.39 (4-6) in 31 dogs.

43 Part I

Weimaraner A medium-sized gray dog, with fine aristocratic features. He should present a picture of grace, speed, stamina, alertness and balance. The average (range) BCS was 4.48 (4-6) in 21 dogs. 3 Whippet A medium size sighthound giving the appearance of elegance and fitness, denoting great speed, power and balance without coarseness. A true sporting hound that covers a maximum of distance with a minimum of lost motion. Should convey an impression of beautifully balanced muscular power and strength, combined with great elegance and grace of outline. The average (range) BCS was 4.08 (3-5) in 28 dogs.

Discussion

The incidence of overweight conditions in dogs in the Netherlands is considered to be similar to that in other countries (i.e. 40-50%) (Ricci et al. 2007, Courcier et al. 2010). The lower incidence of overweight (18.6%) in show dogs can be explained by the fact that these dogs need to be in good show condition, which, for most breeds implies being in an ideal body condition. Health problems are associated with obesity, and are suspected to have a low incidence in show dogs according to the low prevalence of obesity (1.1%) found in this study. All show dogs are sexually intact, which eliminates the effect of neutering on obesity in this population of dogs. Most of the show dogs were young to middle aged, therewith limiting the age effect on obesity in this population of dogs. We did not investigate the other risk factors of obesity (i.e. type of diet, socio-economic status of the owner, changes in lifestyle, and body condition of the owner) in this population of dogs, which might have influenced the results.

We used BCS because it is not invasive and has a good correlation with more accurate methods like chemical analysis, Dual Energy X-ray Absorptiometry, total body water using D2O and bioelectrical impedance (German et al. 2010, Mawby et al. 2004). The use of BCS as a uniform standard for all breeds has been discussed and the use of different BCS scoring methods for different breeds has been proposed because of the variation in body composition within breeds (Jeusette et al. 2010). However, the BCS did correlate well with the Dual-Energy X-Ray Absorptiometry and bioelectrical impedance measurements in that study, so BCS can be effectively used to determine the amount of fat mass in dogs. Whether a certain percentage of fat mass is ideal for dogs of all breeds remains speculative.

In some breeds the ideal show condition seems to be an overweight dog. This can be explained by their body conformation (Jeusette et al. 2010) and breed standards. Historically dogs were selected for certain purposes and an extra fat layer was beneficial for working in a cold climate or to have a reserve for times of scarcity. This selection of

44 Obesity in show dogs

thrifty genotypes is now predisposing to obesity, because the need is lost and periods of scarcity are no longer to be dealt with in most dogs of these breeds, because they are now kept as companion animals (Prentice et al. 2005). In stead, the abundance of food with high energy density may trigger the thrifty genotypes, and therefore, the dogs of these breeds should be monitored more closely to prevent overweight conditions (van 3 den Berg et al. 2010). “Muscular”, “heavier in bone”, “massive build”, “square and thick set in overall build”, “strong”, “masculine”, “heavy thick-set low-swung body”, “dogs are more massive throughout”, “symmetry and general appearance are decidedly square and cobby”, and “on the whole a bold and valiant figure” are all examples of descriptions in breed standards that may be interpreted as being favorable to overweight conditions. These breed standards may encourage breeders to select for thrifty genes, thereby predisposing towards an obese phenotype. On the other hand “elegance”, “smoothly muscled body”, “graceful” and “athletic” may be favorable for lean body condition. Interestingly, “muscular” is used in most breed standards, although in breeds with lower BCS it is often combined with “elegance”. It warrants firm discussions with breeders and judges in order to come to different interpretations of the standards in order to prevent overweight conditions from being the standard of beauty. A genetic predisposition for overweight conditions in some dog breeds should not be an excuse to ignore treatment or prevention programs in these dogs. It should be the other way round, i.e. these dogs should be monitored more closely for body weight and body condition score to prevent development of obesity with the associated health risks.

Acknowledgements

The author greatly acknowledges the Organization of the Winner dog show in Amsterdam 2010 for giving me the opportunity to score the dogs visiting this show. Mr. H.G. Corbee and Mrs. J.J. Corbee-van der Gulik are acknowledged for preparing the data for statistical analysis.

45 Part I

References • Kealy, R.D., Lawler, D.F., Ballam, J.M., Mantz, S.L., Biery, D.N., Greeley, E.H., Lust, G., Segre, • Armstrong, P.J., Lund, E.M., 1996. Changes in M., Smith, G.K., Stowe, H.D., 2002. Effects body composition and energy balance with of diet restriction on life span and age- 3 aging. Veterinary Clinical Nutrition 3, 83–87. related changes in dogs. Journal of the • Berg, L. van den, Berg, S.M. van den, Martens, American Veterinary Medical Association E.E.C.P., Hazewinkel, H.A.W., Dijkshoorn, 220, 1315-1320. N.A., Delemarre-van de Waal, H.A., Heutink, • Laflamme, D. P., 1997. Developmental and P., Leegwater, P.A.J., Heuven, H.C.M., 2010. validation of a body condition score system Analysis of variation in the melanocortin-4 for dogs. Canine practice 22, 10–15. receptor gene (mc4r) in golden retriever dogs. • Laflamme, D.P., 2005. Nutrition for Aging Animal genetics 41, 557-560. Cats and Dogs and the Importance of • Colliard, L., Ancel, J., Benet. J.J., Paragon, B.M., Body Condition. Veterinary Clinics of North Blanchard, G., 2006. Risk factors for obesity America: Small Animal Practice 35, 713-742. in dogs in France. Journal of Nutrition 136, • Mawby, D.I., Bartges, J.W., d’Avignon, A., 1951S-1954S. Laflamme, D.P., Moyers, T.D., Cottrell, T., • Courcier, E.A., Thomson, R.M., Mellor, D.J. 2004. Comparison of various methods for Yam, P.S., 2010. An epidemiological study of estimating body fat in dogs. Journal of the environmental factors associated with canine American Animal Hospital Association 40, obesity. Journal of Small Animal Practice 51, 109-114. 362-367. • McGreevy, P.D., Thomson, P.C., Pride, C., • German, A.J., Holden, S.L., Morris, P.J., Biourge, Fawcett, A., Grassi, T., Jones, B., 2005. V., 2010. Comparison of a bioimpedance Prevalence of obesity in dogs examined by monitor with dual-energy x-ray Australian veterinary practices and the risk absorptiometry for noninvasive estimation factors involved. Veterinary Record 156, of percentage body fat in dogs. American 695-702. Journal of Veterinary Research 71, 393-398. • Nijland, M.L., Stam, F., Seidell, J.C., 2010. • Ishioka, K., Hosoya, K., Kitagawa, H., Shibata, Overweight in dogs, but not in cats, is related H., Honjoh, T., Kimura, K., Saito, M., 2007. to overweight in their owners. Public Health Plasma leptin concentration in dogs: Effects Nutrition 13, 102-106. of body condition score, age, gender and • Prentice, A.M., Rayco-Solon, P., Moore, S.E., breeds. Research in Veterinary Science 82, 2005. Insights from the developing world: 11-15. thrifty genotypes and thrifty phenotypes. • Jeusette, I., Greco, D., Aquino, F., Detilleux, Proceedings of the Nutrition Society 64, J., Peterson, M., Romano, V., Torre, C., 2010. 153-161. Effect of breed on body composition and • Ricci, R., Gottardo, F., Ferlito, J.C., Stefani, comparison between various methods to A., Ravarotto, L., Andrighetto, I., 2007. Body estimate body composition in dogs. Research condition score (BCS) and metabolic status of in Veterinary Science 88, 227-232. shelter dogs. Italian Journal of Animal Science • Kealy, R.D., 1999. Factors influencing lean 6, 859-861. body mass in aging dogs. Compendium on Continuing Education for the Practicing Veterinarian 21, 34–37.

46 Obesity in show dogs

47 48 Chapter 4

Obesity in show cats

Ronald J. Corbee

Journal of Animal Physiology and Animal Nutrition; DOI 10.111/jpn.12176

49 Part I

Abstract

Obesity is an important disease with a high prevalence in cats. Because obesity is related to several other diseases, it is important to identify the population at risk. Several risk factors for obesity have been described in the literature. A higher incidence of obesity 4 in certain cat breeds has been suggested.

The aim of this study was to determine whether obesity occurs more often in certain breeds.

The second aim was to relate the increased prevalence of obesity in certain breeds to the official standards of that breed.

To this end 268 cats of 22 different breeds investigated by determining their body condition score (BCS) on a 9-point scale by inspection and palpation, at 2 different cat shows.

Overall, 45.5% of the show cats had a BCS >5, and 4.5% of the show cats had a BCS>7. There were significant differences between breeds, which could be related to the breed standards. Most overweight and obese cats were in the neutered group.

It warrants firm discussions with breeders and cat show judges in order to come to different interpretations of the standards in order to prevent overweight conditions in certain breeds from being the standard of beauty. Neutering predisposes for obesity, and requires early nutritional intervention to prevent obese conditions.

50 Obesity in show cats

Introduction

Obesity is an important disease with an incidence of 11.5-27% (Cave et al. 2012, Colliard et al. 2009, Courcier et al. 2012). Because obesity is related to several other diseases in cats (i.e. diabetes mellitus, hepatic lipidosis, osteoarthritis) (German et al. 2006, 2010), it is important to identify the population at risk. Several risk factors for obesity in cats have been described in the literature: male gender, neutering, middle age, living in a single or 4 two-cat household, no dog living in the household, inactivity and confinement indoors, feeding fresh meat or fish, eating a premium or therapeutic food, giving food on a free choice or ad libitum basis, and underestimation by the owners of their cat’s body weight or body condition (Cave et al. 2012, Colliard et al. 2009, Courcier et al. 2012). Breed standards have an impact on the visual appearance of a cat and may also influence body condition. They also influence the owner’s perception of optimal body weight and body condition of their cat. A higher incidence of obesity in certain cat breeds has been suggested (Colliard et al. 2009, Kienzle and Moik 2011). In dogs significant differences between breeds, which could be related to the breed standards were already demonstrated (Corbee 2013). A possible relationship between breed standards and overweight is not yet determined in cats. However, the body weight of purebred client-owned cats was studied (Kienzle and Moik 2011), as well as the incidence of overweight conditions in these cats. Significant differences between neutered Norwegian Forest Cats and neutered Siamese/Oriental Shorthair Cats on the incidence of overweight conditions were found. The data on body condition score (BCS) per breed were not presented, because the primary aim of that study was to determine optimal body weight in pure-bred cats. According to Colliard et al. (2009) purebred cats have a lower risk for obesity, and Domestic Shorthair and crossbred cats are at risk for obesity. The aim of this study is to determine whether there are differences between the BCS of show cats of a variety of breeds, and to relate these differences to the breeding standards. Our hypothesis was that the BCS of breeds that are to be powerful and muscular, as defined by the breed standards (i.e. British shorthair), is higher, whereas other breeds are expected to have a lower BCS because they are to be graceful and lithe, as defined by the breed standards (i.e. Abyssinian).

Materials, methods and animals

We determined BCS of 268 cats of 22 different breeds on a 9-point scale (Laflamme et al. 2005) by inspection and palpation at two different cat shows. BCS of all cats was performed by the same board certified nutritionist (R.C.). Before entering the cat show, all cats need to be checked by a veterinarian. The cats were randomly assigned to a consultation room, at 1 of which the BCS was performed by the nutritionist, so not all the cats that visited the cat shows were scored. The assistant of the cat show, as well as the cat owners visiting the cat show were unaware of the performed study before entering the consultation room. Cats that were on both shows, were only included once. Breed standards were obtained from the American Cat Fanciers Association, and were investigated for descriptions that may predispose cats to become overweight or obese.

51 Part I

Statistical analysis A multi analysis of variance was performed to determine differences within breeds and to determine possible influences of sex, neuter status and age on BCS. A Kolmogorov- Smirnov test was performed for normality. Because not all BCS data were normally distributed, a non-parametric test was used. A Kruskal-Wallis test was performed to demonstrate differences between breeds. The scores of the breeds were tested against 4 the mean value of all show cats. A p-value of 0.05 was set as the level of significance. Overweight was defined as a BCS >5 on a 9-point scale. Obesity was defined as a BCS >7 on a 9-point scale (Laflamme et al. 1997). Age groups were defined as kittens (<1 year) or adults (1-5 years). Cats older than 5 years of age were excluded from the study. Based on a power-analysis a breed group should consist of at least 5 cats in order to demonstrate significant differences with an alpha of 0.05 and a beta of 0.10. Breeds with lower numbers of cats were excluded from the results, but were included in the overall calculation of the prevalence of overweight conditions in show cats. Because most data were normally distributed, data are shown as mean±SD.

Results BCS Overall, 45.5% of the show cats were overweight, and 4.5% of the show cats were obese (Table 1). The average BCS of all investigated show cats (n=268) was 5.55±0.93 on a 9-point scale (Table 2). Most overweight cats were in the neutered group (neutered males 6.63±0.97, neutered females 6.55±1.04). There were no significant differences between males and females within breeds, however, in total; males had a significantly higher BCS (5.81±1.02) compared to females (5.29±0.75) (p<0.05). Most lean cats were kittens with no significant difference between sexes (males 5.00±0.00, females 5.19±0.40). In adult cats the BCS was 5.48±0.83 for intact males, and 5.18±0.64 for intact females (p=0.183). Significantly higher BCS compared to the average BCS of all show cats was scored in: British shorthair (5.92±0.95), Norwegian Forest Cat (5.86±0.88), and Persians (6.27±1.10). Significantly lower BCS compared to the average BCS of all show cats was scored in Abyssinian (4.43±0.53), Cornish Rex (4.40±0.55), Devon Rex (4.60±0.55), Oriental Shorthair (4.67±0.52), and Sphynx (4.40±0.55).

Table 1: Percentage of adult show cats being overweight (BCS >5) or obese (BCS >7)

BCS >5 BCS>7 Total show cats 45.5 4.5 Intact adults 35.2 0 Males 43.6 0 Females 28.7 0 Neutered adults 83.6 23.1 Males 90.2 24.4 Females 81.8 18.1

52 Obesity in show cats

Table 2: Mean Body Condition Score (BCS) of show cats of different breeds

Description Number of cats Mean BCS ±sd Age and gender Total show cats 268 5.55±0.93 Males 135 5.81±1.02 * 4 Females 133 5.29±0.75 * Kittens Males 7 5.00±0.00 Females 21 5.19±0.40 Intact adults Males 94 5.48±0.83 Females 122 5.18±0.64 Neutered adults Males 41 6.63±0.97 Females 11 6.55±1.04

Breeds Cornish Rex 5 4.40±0.55 ** Sphynx 5 4.40±0.55 ** Abyssinian 7 4.43±0.53 ** Devon Rex 5 4.60±0.55 ** Oriental Shorthair 6 4.67±0.52 ** Siamese 13 4.69±0.48 Domestic Shorthair 7 5.29±0.49 Maine Coon 62 5.52±0.92 Average 268 5.55±0.93 **,*** Ragdoll 17 5.59±0.71 Burmese 10 5.60±0.70 Birman 20 5.70±0.47 Norwegian Forest Cat 54 5.86±0.88 *** British Shorthair 37 5.92±0.95 *** Persian 11 6.27±1.10 ***

*Significant difference between males and females (p<0.05) **Significantly lower BCS compared to average BCS (p<0.05) ***Significantly higher BCS compared to average BCS (p<0.05)

53 Part I

Breed standards Breed standards (i.e. phrases from the general appearance) of the breeds with significantly higher body condition score British Shorthair The British Shorthair is a medium to large almost square cat. It is a sturdy cat, well knit and powerful with a broad, rounded chest. The head is broad with well rounded contours 4 when viewed from any angle, with full cheeks, giving a chubby appearance, with a short broad nose. Overall appearance of this breed with its short bull neck, standing on strong muscular legs and well rounded paws, is that of a solid muscular cat, with no fat on its body, pleasing to the eye and very amenable to handling.

Norwegian Forest Cat The Norwegian Forest Cat is a sturdy cat with a distinguishing double coat and easily recognizably body shape. It is a slow maturing breed, attaining full growth at approximately five years of age. Body: moderate in length, substantial bone structure, with a powerful appearance showing a broad chest and considerable girth without being fat. Great depth of flank. Males are large and imposing; females may be smaller and more refined.

Persian The ideal Persian should be a well proportioned, medium to large cat, giving the impression of elegant robust power. The head is massive with small rounded ears, wide set eyes and a short nose presenting a sweet expression. In profile, the forehead, the nose and the chin appear to form a perpendicular line. The body is short and cobby, well balanced with heavy boning and short legs. Tail is in proportion to the body. It is recognized in a variety of colors and patterns.

Breed standards of the breeds with significantly lower body condition score Abyssinian The overall impression of the ideal Abyssinian is a medium to large cat, regal in appearance. Males are proportionately larger than females. The Abyssinian shows firm muscle tone and is lithe and panther-like in activity showing a lively interest in all surroundings. The Abyssinian is of sound health and vigor, physically well balanced, and amenable to handling.

Cornish Rex The Cornish Rex is characterized by a firm, muscular body, well toned with no evidence of obesity or emaciation. Viewed as a whole, it is well knit, alert, standing high on its legs with each part in good proportion. All contours are gently curved. It is a medium sized cat, slender and fine boned with a full, deep rib cage. The trunk of the body follows an upward curve of the backbone forming a “tuck-up” accented by its rounded hips. The most unique aspect of the breed is its coat with deep even waves over the entire body, legs and tail.

54 Obesity in show cats

Devon Rex The Devon Rex is a breed of unique appearance. Its large eyes, moderately short muzzle, huge low set ears and prominent cheekbones, create a characteristic Pixie look. A cat of medium fine frame, the Devon is well covered with soft wavy fur; the fur is of a distinctive texture. The Devon is alert and active and shows a lively interest in its surroundings. Mature Males may be expected to be significantly larger than females, and should not be faulted based on size. 4

Oriental shorthair The ideal Oriental Shorthair cat should be graceful, svelte, muscular and fine-boned. It should be nicely proportioned and well balanced, with long tapering lines. The cat should be in excellent physical condition with clear, bright eyes.

Sphynx The Sphynx appears to be a hairless cat, although it is not truly hairless. The cat should not be small or dainty. Males may be up to 25% larger as long as proper proportions are maintained. The cat should not be penalized or faulted for any evident scars, scratches or blemishes on its body. The Sphynx is sweet-tempered, lively, intelligent and above all, amenable to handling.

Discussion

The overall incidence of obesity (4.5%) and overweight (45.5%) in show cats is lower compared to the incidence of obesity (11.5-27%) and overweight (63%) in cats from previous studies (Cave et al. 2012, Colliard et al. 2009, Courcier et al. 2012). Most of the cats in our study were intact cats, which had to be in perfect show condition, while neutering is one of the major risk factors for overweight (OR 3.55-9.3) (Cave et al. 2012, Colliard et al. 2009, Courcier et al. 2012). This is also demonstrated by the highest prevalence of overweight show cats in the neutered group.

Significant differences between cat breeds were demonstrated, which could be related to the breed standards. In the standards of the more obese prone breeds; terms like chubby, cobby, sturdy, square, round, powerful, muscular, broad chested, short bull neck are mentioned, which might favor overweight conditions. In the standards of lower BCS breeds; terms like graceful, lithe, balanced, fine-boned, svelte, nicely proportioned are mentioned. Surprisingly, Maine Coon also have obese prone descriptions in the breed standards, but did not have a significantly higher BCS compared to other cat breeds, however the average BCS in Maine Coons was well above 5 (5.52±0.92). Interestingly, in the breed standard of the Norwegian Forest Cat it is mentioned that the cats should not be fat, while 55% of them were overweight.

Body condition score was used to be able to score, uniformly, a large number of cats of a variety of breeds. Despite the fact that not all the cats were scored (55% were scored), the

55 Part I

large number of cats included is still regarded as representative for the total population of show cats in the Netherlands. BCS was used because it is non-invasive and has a good correlation with more accurate methods like Dual Energy X-ray Absorptiometry (Bjornvad et al. 2011). Furthermore it has a better correlation with the percentage of body fat compared to other non-invasive measurements such as body weight or girth measurement (Bjornvad et al. 2011). Body condition scoring in cats has more inter- 4 observer variation compared to BCS in dogs (Bjornvad et al. 2011). Most veterinarians tend to underestimate BCS in cats, as is noted by the correlation study with DEXA, where cats with a body fat percentage of 31% were still considered to have a BCS of 5, while this body fat percentage should correspond with a BCS of 6 (Bjornvad et al. 2011).

The results correspond with the data of Kienzle and Moik (2011), although a higher incidence of a BCS>5 in intact cats was found (35.2%: 43.6% of intact males and 28.7% of intact females compared to 7% of intact males and 3% of intact females), however, none of the intact males and intact females in our study had a BCS >7. In neutered cats the incidence of overweight was also higher (90.2% of neutered males and 81.8% of neutered females, compared to 50% of neutered males, 38% of neutered females), a BCS>7 was scored in 24.4% of neutered males and 18.1% of neutered females. No significant differences in BCS of breeds in the neutered group were demonstrated, while in the study of Kienzle and Moik neutered Norwegian Forest cats had a higher incidence of overweight, whereas neutered Siamese/Oriental shorthairs had a lower incidence of overweight. We did not study cats without pedigree. In the study of Kienzle and Moik cats without pedigree 51% of neutered males and 26% of neutered females were scored overweight. We found no significant differences between males and females within breeds, but this might be to a lack of power, due to the lower number of cats per breed. Overall, male cats had a significantly higher BCS (5.81±1.02) compared to females (5.29±0.75). This corresponds with other studies that mentioned male gender as a risk factor for obesity in cats (Cave et al. 2012, Colliard et al. 2009, Courcier et al. 2012). Because of time issues, no complete nutritional assessment was performed (Baldwin et al. 2010). We excluded cats > 5 years of age because only a few cats were presented in this age category. Furthermore, from this age a decrease in energy requirements has been demonstrated in cats over 5 years of age, which predisposes these cats for obesity (Laflamme et al. 2005). When cats are over 10-12 years of age, the energy requirements tend to increase again (Cupp et al. 2004).

Colliard et al. (2009) found an OR for being overweight of 0.73 for crossbred, and 0.3 for purebred, compared to Domestic shorthair (OR=1.0), although they only included 33 purebred cats (of 10 breeds, only 3 breeds with >4 cats: 15 Persian, 6 Birman and 6 Siamese) in a population of 385 cats. Longhair cats had an OR for being overweight of 0.26 (n=95) compared to shorthairs (OR=1.0; n=214). They did not state if there were differences between breeds. The scoring used in this study was a 5-point BCS, which makes it more difficult to compare their data with our results.

56 Obesity in show cats

Courcier et al. (2012) found no breed with an increased risk of obesity, but again, only small numbers of purebred cats (n=120 of 16 breeds, only 5 breeds with >4 cats: 43 Persian, 26 Siamese, 10 British Shorthair, 7 Bengal, 6 Ragdoll compared to the total study population (n=3217). They found no significant differences in BCS between intact males and intact females, or between neutered males and neutered females. The cat population investigated by Courcier et al. (2012) were admitted to veterinary charity hospitals, which, according to them, implies owners of the lower socio-economic class, that have 4 less money to spend on cat food and have there cats walking outside more compared to cats in other households. The overall BCS>5 in their study was very low (11.5%) compared to our study (45.5%).

The population of show cats is not representative for the total population of purebred cats. We did not weigh the cats and had no information on housing and feeding practice. In the Netherlands, most purebred cats are housed solely indoors, which might be an explanation for the high incidence of overweight conditions in show cats. Furthermore, Kienzle and Moik (2011) stated that breeders like the bigger cats of certain breeds more, compared to the smaller individuals, which may select for genes that are associated with overweight.

When compared to the study in dogs (18.6% overweight, and 1.1% obese) (Corbee 2013), the incidence of obesity in show cats (45.5% overweight, and 4.5% obese) is higher. This may be explained by the fact that cat breeds are not developed for work properties, but mainly for exterior. Also, in show dogs no neutered dogs were presented, but the incidence of overweight in intact show cats (35.2% overweight) is still higher compared to intact show dogs (18.6%). The incidence of obesity (i.e. BCS >7 on a 9-point scale) in intact show cats is lower (0.0%) compared to show dogs (1.1%). Furthermore, less cats of the breeds with a lower incidence of overweight were present at the two cat shows compared to cats of the breeds with a higher incidence of overweight.

Owners were not asked to give their opinion about the body composition of their cats. We know from previous studies that most owners underestimate the body condition of their cat, which is known as a risk factor for obesity (Cave et al. 2012, Colliard et al. 2009, Courcier et al. 2012). The image of what is considered a normal cat is actually an overweight cat. This is similar to findings in certain dog breeds (Corbee 2013).

Neutering is confirmed to be a risk factor for obesity as 83.6% of neutered cats had a BCS>5, versus 35.2% of intact cats. Neutering reduces energy requirements (Alexander et al. 2011, Mitsuhashi et al. 2011), while increasing appetite in cats (Cave et al. 2007). It is important for veterinarians to firmly discuss this with the owners, and give specific nutritional recommendations to prevent neutered cats from developing overweight or obese conditions.

57 Part I

Conclusion

There were significant differences between breeds, which could be related to the breed standards. It warrants firm discussions with breeders and cat show judges. This different interpretation is needed to come to different interpretations of the standards in order to prevent overweight conditions in certain breeds from being the standard of beauty. 4 Neutering predisposes cats for obesity, and requires early nutritional intervention to prevent occurrence of obese conditions in these cats.

58 Obesity in show cats

References • Courcier, E.A., Mellor, D.J., Pendlebury, E., Evans, C., Yam, P.S., 2012. An investigation into • Alexander, L.G., Salt, C., Thomas, G., the epidemiology of feline obesity in Great Butterwick, R., 2011. Effects of neutering Britain: results of a cross-sectional study of on food intake, body weight and body 47 companion animal practices. Veterinary composition in growing female kittens. British Record 171, 560-564. Journal of Nutrition 106, S19-S23. • Cupp, C., Perez-Camargo, A., Patil, A., Kerr, W., 4 • Baldwin, K., Bartges, J., Buffington, T., 2004. Long-term food consumption and body Freeman, L.M., Grabow, M., Legred, J., Ostwald weight changes in a controlled population of Jr., D., 2010. AAHA nutritional assessment geriatric cats. Compendium on Continuing guidelines for dogs and cats. Journal of the Education for the Practicing Veterinarian 26 American Animal Hospital Association 46, (Suppl. 2A), 60. 285-296. • German, A.J., 2006 The growing problem of • Bjornvad, C.R., Nielsen, D.H., Armstrong, obesity in dogs and cats. Journal of Nutrition P.J., McEvoy, F., Hoelmkjaer, K.M., Jensen, 136, 1940s-1946s. K.S., Pedersen, G.F., Kristensen, A.T., 2011. • German, A.J., Ryan, V.H., German, A.C., Wood, Evaluation of a nine-point body condition I.S., Trayhurn, P., 2010. Obesity, its associated scoring system in physically inactive pet cats. disorders and the role of inflammatory American Journal of Veterinary Research 72, adipokines in companion animals. The 433-437. Veterinary Journal 185, 4-9. • Cave, N.J., Allan, F.J., Schokkenbroek, S.L., • Kienzle, E., Moik, K., 2011. A pilot study of Metekohy, C.A.M., Pfeiffer, D.U., 2012. A cross- the body weight of pure-bred client-owned sectional study to compare changes in the adult cats. British Journal of Nutrition 106, prevalence and risk factors of feline obesity S113-115. between 1993 and 2007 in New Zealand. • Laflamme, D.P., 2005. Nutrition for Aging Preventive Veterinary Medicine 107, 121-133. Cats and Dogs and the Importance of • Cave, N.J., Backus, R.C., Marks, S.L., Klasing, Body Condition. Veterinary Clinics of North K.C., 2007. Oestradiol and genistein reduce America: Small Animal Practice 35, 713-742. food intake in male and female overweight • Mitsuhashi, Y., Chamberlin, A.J., Bigley, cats after gonadectomy. New Zealand K.E., Bauer, J.E., 2011. Maintenance energy Veterinary Journal 55, 113-119. requirement determination of cats after • Colliard, L., Paragon, B.M., Lemuet, B., Bénet, spaying. British Journal of Nutrition 106, J.J., Blanchard, G., 2009. Prevalence and risk S135-S138. factors of obesity in an urban population of healthy cats. Journal of Feline Medicine and Surgery 11, 135-140. • Corbee, R.J., 2013. Obesity in show dogs. Journal of Animal Physiology and Animal Nutrition 97, 904-910.

59 60 Chapter 5

Chardin 1728 Still Life with Cat and Fish The effect of dietary long- chain omega-3 fatty acids supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

Ronald J. Corbee Mieke M.C. Barnier Chris H.A. van de Lest Herman A.W. Hazewinkel

Journal of Animal Physiology and Animal Nutrition 2013; Volume 97, Issue 5, Pages 846-853

61 Part I

Abstract

The aim of this randomized, double-blinded, placebo controlled cross-over designed study was to demonstrate the clinical effect, registered by a survey, of a 10-week period of omega-3 fatty acids supplementation of the diet (1.53g EPA and 0.31g DHA, both per 1000kcal ME, equivalent to the complete diet) of 16 cats with radiologically documented, naturally occurring osteoarthritis, in 5 comparison with a 10-week period of supplementation with corn oil (0.00g EPA and 0.00g DHA, both per 1000kcal ME). Cats on the fish oil revealed higher activity level (p=0.07), more walking up and down the stairs (p=0.07), less stiffness during gait (p=0.03), more interaction with the owner (p=0.07) and higher jumps (p=0.03) compared to corn oil supplementation.

In conclusion: supplementation of long chain omega-3 polyunsaturated fatty acids changes the owner’s perception of some aspects of behavior and locomotion in cats with naturally occurring osteoarthritis.

62 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

Introduction

Osteoarthritis (OA) in cats is getting more awareness in the recent years by researchers and veterinarians. OA is initiated by trauma of joint cartilage (secondary OA) and/or shortening of the proteoglycans with concurrent increase of advanced glycation end products (primary OA). Pathogenesis of OA is characterized by a chronic inflammatory state resulting in an increase in matrix metalloproteinases (MMPs) that perpetuates the cycle of destruction of cartilage and inflammation (Allan 2000, Anderson et al. 2004, Burns 2006, Kerwin 2010, Slingerland et al. 2011). OA is a common finding in older cats, 5 with a prevalence of 22-72% in cats >6 years of age (Clarke et al. 2005, Godfrey 2005, Hardie et al. 2002, Lascelles 2008). The prevalence and severity of OA increase with age (Slingerland et al. 2011). The vertebrae, elbows, hips, shoulders and tarsi are the most commonly affected joints (Slingerland et al. 2011). The behavioral changes of cats with OA are decreased mobility, less grooming, and defecating outside the litter box (Clarke et al. 2005, Clarke and Bennett 2006, Hardie 1997, Rothschild et al. 1998, Slingerland et al. 2011). The first signs of OA are mostly subtle, so OA is commonly recognized at a progressive state of disease and often remains undiagnosed and untreated in the earlier stages (Godfrey 2005, Lascelles 2010, Rychel 2010). It may be difficult to detect painful OA in cats, since cats tend to adapt their mobility to their abilities to avoid pain (Taylor and Robertson 2004). These behavioral adaptations are mistakenly considered normal for an aging cat by the owner, and are not related to OA. Nevertheless, Zamprogno (2010) and Slingerland et al. (2011) demonstrated that activity related behavior, as noted by cat owners, is significantly different between healthy cats and cats with signs of OA- associated pain. Treatment of cats suffering from OA will include guided weight reduction (Scarlet and Donoghue 1998), non-steroid anti-inflammatory drugs (NSAIDs) (Clarke and Bennett 2006, Gunew et al. 2008) and nutritional management (Gück 2009). Scarlett and Donoghue (1998) demonstrated a 4.9 times increased risk of lameness due to OA in obese cats versus lean cats that were admitted to veterinary hospitals and followed up during 4.5 years. Obese cats may have more severe OA and OA associated pain compared to lean cats, due to the chronic release of pro-inflammatory cytokines by adipose tissue, and the release of leptin from chondrocytes, as is demonstrated in men and research animals (obese/obese mouse) (Simonpoulou et al. 2007). Weight loss improved locomotion in overweight dogs with OA (Gück 2009, Hardie 1997, Laflamme 2005, Marshall et al. 2010, Rychel 2010). Nutritional management of OA may further include modulation of the inflammatory response. Nutrients that are claimed to modulate the inflammatory response are long chain omega-3 fatty acids (Gück 2009). Omega-3 fatty acids sort their effect by altering the cell wall fatty acid composition (Cao et al. 2006), resulting in a competition for cyclo-oxygenase enzymes between eicosapentaenoic acid (EPA), which promotes the anti-inflammatory mediators (i.e. Leukotriene B5, Prostaglandin E3 and Thromboxane 3 series), and arachidonic acid (AA), which promotes the pro-inflammatory mediators (i.e. Leukotrienes B4, Prostaglandin E2 and Thromboxane 2 series). The increase of anti-inflammatory mediators results in decreased production of MMPs as well as increased production of tissue inhibitors of MMPs. The effects of omega-3 fatty acid

63 Part I

supplements on OA have been studied in different species, including man (Cao et al. 2006) and dogs (Fritsch et al. 2010, Frost-Christensen et al. 2006, Roush et al. 2010a, Roush et al. 2010b). Omega-3 fatty acids have been proposed to have a beneficial effect on OA in cats (Gück (2009), and this has been investigated in several studies. They have anti-arrhythmic effects (Freeman et al. 2010) and also have an anti-inflammatory action on bowel disease (Trepanier et al. 2009). Further, a multicomponent therapeutic diet with elevated EPA and DHA levels has been shown to relieve OA associated pain in cats, when compared to a control food, (Lascelles et al. 2010). These effects were registered by accelerometry and 5 behavioral changes. However, it is unclear whether these effects were due to increased EPA and DHA levels or due to concomitant additional nutritional modifications, such as green-lipped mussel extract and glucosamine/chondroitin sulphate (Beale 2004).

Material and methods Animals Owners of 35 cats, referred to the University department all with radiographically confirmed OA as part of a previously reported study (Slingerland et al. 2011), were invited to participate in this study. Owners of 24 cats were willing to participate in our study. Radiographs of both shoulders (mediolateral [ML]), elbows (ML), carpi (dorsopalmar) including front leg digits, coxofemoral joints (ventrodorsal), stifles (ML and craniocaudal), and tarsi (ML) were evaluated. For each joint, OA was graded as absent, minimal, moderate or severe, according to Hardie et al. (2002). The inclusion criteria were >8 years of age, moderate OA (according to Hardie et al. (2002)) in one or more joints, being on a commercially available dry cat food, and the willingness of the owners to cooperate in this study (i.e. daily supplementation of the daily ration with an exact amount of oil). Exclusion criteria were the use of NSAIDs and the use of (dietary) supplements from 2 weeks before the start of the study and during the study.

Questionnaire A questionnaire based on the previous study in these cats was used (Slingerland et al. 2011) completed with questions on behavior and activity, according to Zamprogno et al. (2010). The questionnaire was tested prior to being used by interviewing cat owners in the waiting room of our clinic and necessary adaptations were made. The questionnaire included only closed questions with predetermined and pre-coded categories (Appendix 1). The same researcher (M.M.C.B.) had telephone contact with the owner prior to the study to ask for cooperation and completion of the survey, and after both supplementation periods, without knowing if the cat just completed a 10-week period on oil “A” or oil “B”.

64 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

Table 1: Fatty acid composition of the maintenance cat foods

Cat number EPA+DHA Total n-6 Total n-3 n:6-n-3 1 0.26 n/a n/a n/a 2 0.05 n/a n/a n/a 3 0.42 3.22 0.80 4.00 4 0.05 0.05 n/a n/a 5 0.05 0.05 n/a n/a 5 6 0.20 3.80 0.54 7.00 7 0.32 4.62 0.81 6 8 0.26 3.45 0.60 6 9 n/a 3.22 0.39 8 10 n/a 3.22 0.39 8 11 n/a 3.22 0.39 8 12 0.05 n/a n/a n/a 13 0.70 2.88 1.08 3 14 0.05 n/a n/a n/a 15 n/a 2.78 0.70 4 16 0.05 n/a n/a n/a 17 0.26 3.45 0.60 6 18 n/a 2.78 0.70 4 19 0.26 3.45 0.60 6 20 0.05 n/a n/a n/a

Data in % on dry matter basis, obtained from manufacturer EPA = eicosapentaenoic acid DHA = docosahexaenoic acid n-6 = omega-6 polyunsaturated fatty acids n-3 = omega-3 polyunsaturated fatty acids n/a = not available

65 Part I

Supplements The supplements (Catoils® joint, and corn oil) were provided by Nutriceuticoils®, Belgium. The test and control supplement had an equal appearance (i.e. corn oil with fish smell and fish oil (mainly derived from Anchovy (Engraulidae) and Sardine (Clupeidae)) in identical bottles) and were indicated by a letter “A” or “B” on the 100mL bottle. The owners were advised to provide the oil in a dosage of 1ml per 5kg body weight daily to their cats, (dosage syringe was provided) added to the declared given maintenance food (i.e. a commercially available dry cat food, fed dry, which meets NRC requirements for all 5 life stages) (NRC 2006). The key was broken after statistical analyses of all study results. The amount of oil made available to the owner was sufficient for the test period and did not contain an extra volume to prevent errors of long-term administration of the product during the next test period. Oil “A” contained 0mg/ml ETA, 0mg/ml EPA and 0mg/ml DHA. Oil “B” contained 15mg/ml ETA, 500mg/ml EPA and 100mg/ml DHA. Vitamin E (alpha- tocopherol 10mg/ml) was added to both formulations to prevent rancidity. The fatty acid compositions of the provided maintenance cat foods, as declared by the manufacturer, are shown in table 1. The fatty acid composition of oil “A” and oil “B”, as determined by an independent laboratory (Napro Pharma AS, Norway) using method Ph.Eur.2.4.29 (chromatography) is shown in table 2.

Study protocol Before treatment with either oil “A” or oil “B”, a questionnaire was completed by the owners. The cats were then randomly given either oil “A” or oil “B” in a randomized matter for 10 weeks. After this 10-week period the questionnaire was completed by the owners. Then another 10-week period of supplementation with the other oil was performed, followed by the final questionnaire. No wash-out period was introduced between the 2 periods.

Control group From the Department’s colony, 20 healthy cats (all neutered), were used to investigate the blood fatty acid concentrations before and after a 10-week period on one of the two investigated supplements, given randomly (i.e., daily supplementation with oil “A” or oil “B”, respectively). The animals were fed a dry diet which was supplemented daily with either oil “A” or oil “B”. Blood samples were taken, before and after the 10-week period, by jugular venipuncture and collected in heparinised tubes. The blood was immediately centrifuged and the plasma harvested and stored at -20°C until assay. These samples served to determine the fatty acid profile of the cholesteryl ester fraction of plasma (see below). The investigators were blinded to the oils the cats were given during this 10-week period.

Analysis of plasma fatty acid composition Cholesteryl esters (CEs) in plasma of day 0, day 70 and day 140 were isolated by a modified “Bligh & Dyer” extraction according to Retra et al. (2008), followed by a solid- phase extraction method according to Hamilton and Comai (1988). Thereafter the

66 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

Table 2: Fatty acid composition of the oils

Fatty acid Fish oil g/100g Corn oil g/100g Omega-6 4.60 51.80 C18:2 (LA) 1.30 51.80 C20:4 (AA) 2.10 0.00

Omega-3 63.60 1.00 C18:3 (ALA) 0.60 1.00 5 C20:4 (ETA) 1.60 0.00 C20:5 (EPA) 43.90 0.00 C22:5 (DPA) 2.20 0.00 C22:6 (DHA) 11.30 0.00

Other C16:0 (Sat.)(PA) 0.60 10.23 C16:1 (n-7)(PO) 0.20 0.10 C18:0 (Sat.)(ST) 4.60 1.80 C18:1 (n-9)(OL) 9.70 28.60

Total PUFA 69.50 52.80 Total MUF 16.40 29.50 Total SAF 6.40 12.80 Total unknown FA 2.50 0.00

Data in % on dry matter basis, obtained from Napro Pharma AS, Norway

LA = linoleic acid ST = stearic acid AA = arachidonic acid OL = oleic acid ALA = alpha linolenic acid PUFA = polyunsaturated fatty acids ETA = eicosatetraenoic acid MUF = monounsaturated fatty acids EPA = eicosapentaenoic acid SAF = saturated fatty acids DPA = docosapentaenoic acid FA = fatty acids DHA = docosahexaenoic acid Peroxide Value maximum 5 meq/kg PA = palmitic acid Anisidine Value maximum 15 meq/kg PO = palmitoleic acid

67 Part I

CEs were saponified according to the modified method described by Kates (1986), where petroleum ether was replaced by hexane. Polyunsaturated fatty acid analysis was performed by High Pressure Liquid Chromatography / mass spectrometry (HPLC/ MS) according to Retra et al. (2008) in which the Synergi 4 µm MAX-RP 18A column was replaced by a Kinetex 2.6 µm C18 100A column (150 x 3 mm; Phenomenex, CA, USA). Internal standards were used for comparison.

Statistical analysis 5 A Kolmogorov Smirnov test was used to test for normality. Because most values were not normally distributed, non-parametric tests were used. Because no wash-out period was introduced, a Grizzle test for carryover effect was performed on all parameters tested. A p-value of <0.10 was considered significant, because the study design is a 2-way crossover. A Mann-Whitney-U-test has been performed to determine the differences between treatment groups regarding answers to the questions in the client-owned cats, followed by a Bonferroni correction for multiple comparisons. A Mann-Whitney-U-test has been performed to determine the differences between fatty acid patterns of the cholesteryl fraction of plasma in the cats from the Department’s colony.

A p-value of <0.05 was considered significant. A p-value of <0.10 was set as trend. All statistical analyses were done using SPSS 16.0 for Windows (SPSS Inc. Chicago, IL, USA). All studies described in this article were approved by the Committee of Experimental Committee Use of Animals of Utrecht University and performed with written consent from the owners of the investigated 24 privately owned cats and from the owner of the 20 cats from the Department’s colony.

Results

From the 35 owners contacted, 24 owners approved. Three cats were lost to follow up for different reasons (1 owner didn’t want to participate anymore after 2 weeks, 1 cat didn’t like the taste of the oil “A” from day 1 on and 1 cat started vomiting when fed the oil “B” on day 3) so we completed the study with 21 cats 11 of which first received oil “A” and 10 first received oil “B”. The 20 client-owned cats included 2 Abyssinians, 16 Domestic shorthairs and 2 Persians. The age was 13.0±2.9 years. All the cats were neutered, 11 cats were female and 9 were male. Most cats accepted the oil well, 2 cats occasionally vomited and 3 cats didn’t like the taste of the oils, but the owners managed to provide the oils orally to the cat by a syringe followed by a meal. After 2 days one cat died and was censored. This cat was on oil “A”. After the first 10 week period 4 cats were lost to follow up, so the first period was completed with 20 cats and the second period was completed with 16 cats. None of the included cats were in need of any medication during the study period. The data of the 4 cats that did not complete the second period were censored.

The owner’s perception of cat’s behavior and well-being, in cats on the oil “A” compared to the oil “B”, has been evaluated: the results are shown in table 3. After a 10-week period

68 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

Table 3: Questionnaire results

Question p-values Mann-Whitney

3. Time grooming 0.564 4. Defecation outside litter box 1.000 5. Change in usage of litter box 1.000 6. Play with other pets 0.381 5 7. Change in activity level 0.070 8. Differences in jumping on objects 0.564 9. Change in jumping height 0.030 11. Change in walking up the stairs 0.070 12. Change in walking down the stairs 0.070 13. Stiffness during gait 0.030 14. Change in aggression 0.780 15. Change in satisfaction 0.564 16. Change in interaction with people 0.070 17. Change in petting time 0.780 18. Change in interaction with cats 1.000 19. Changes in self-grooming moments 0.138 20. Changes in self-grooming time 0.138

Questions are shown in appendix 1 Mann-Whitney compared 10-week period on oil “A” versus oil “B” p<0.05 is set as the level of significance and significant results are shown in bold p<0.10 is set as trend and results are shown in italic

69 Part I

on oil “B”, the cats revealed higher activity level (p=0.07), more walking up and down the stairs (p=0.07), less stiffness during gait (p=0.03), more interaction with the owner (p=0.07) and higher jumps (p=0.03) when compared with a 10-week period on oil “A”. Compared with the study of Slingerland et al. (2011) we did not demonstrate a significant difference in defecation outside the litter box, because none of the cats in this study showed this clinical sign before, during or after the study periods.

There was improvement of behavior during play with other pets, jumping on objects, 5 moments of grooming, and grooming time with the use of both oils, and did not reach statistical significance between groups (p=0.381, p=0.564, p=0.138 and p=0.138, respectively). No carryover effects were observed (p>0.10). The 20 cats of the Department’s colony were all Domestic Shorthairs. The age was 6.5±3.3 years. All cats were neutered, 10 cats were male and 10 cats were female.

Administration of oil “B” was reflected in the fatty acid pattern of the cholesteryl fraction of plasma of the 10 cats by significant increase (before versus after) in eicosatrienoic acid (p=0.041), eicosatetraenoic acid (p=0.041), eicosapentaenoic acid (p=0.041), docosapentaenoic acid (p=0.049) and docosahexaenoic acid (p=0.041). The 10 cats on oil “A” demonstrated no significant differences in the fatty acid pattern of the cholesteryl fraction of plasma after 10 weeks of supplementation. At the conclusion of the investigation the key was disclosed, where oil “A” appeared to be the control oil and oil “B” the test oil (table 2).

Discussion

The aim of this study was to investigate the effect of a 10-week period on omega-3 fatty acids supplementation (1.53g/1000kcal ME EPA and 0.31g/1000kcal ME DHA) on the owner’s perception of behavior and locomotion of cats with known naturally occurring OA, in comparison with a 10-week period on corn oil supplementation. We used a 10- week period, as described in previous studies in man being long enough to measure effects of long-chain omega-3 polyunsaturated fatty acids (Cao et al. 2006, Hansen et al. 1998, Masson et al. 2007). A larger dose or a longer period of the fish oil supplementation could possibly have given an even better result. However, Blonk et al. (1990) showed that, in humans, doses over ~1.2 g DHA/day (given as fish oil) saturates the plasma DHA concentration and further increase of given DHA increases the plasma concentration only incrementally, suggesting a certain maximum. We did not introduce a wash-out period because this would extend the study period and may have led to decreased compliance of the cat owners. This might have caused carryover effects, although we could not demonstrate them in our results. Compared with studies in dogs by Lascelles et al. (2010) (i.e. 1.88g EPA+DHA per 1000 kcal ME), Roush et al. (2010a, 2010b) (i.e. 1.00g DHA per 1000kcal ME) and Fritsch et al. (2010) (i.e. 0.34; 0.90 and 1.35g DHA per 1000kcal ME) the DHA content in our study is at the lower end (i.e. 0.31g DHA per 1000kcal ME). The EPA

70 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis content in our study was at the higher end (i.e. 1.53g EPA per 1000kcal ME) compared to the other studies (i.e. 1.88g EPA and DHA per 1000kcal ME; 0.41g EPA per 1000kcal ME and 0.45; 1.1 and 1.6g EPA per 1000kcal ME). In this study we aimed at a certain absolute amount of EPA and DHA originating from fish oil, rather than omega-3:omega-6 ratio’s, since recent studies in man demonstrated that the absolute amounts of EPA and DHA are more important (Zainal et al. 2009). The test oil “B” also contained 15mg/mL ETA which may also contribute to clinical improvement as being a precursor of EPA and DHA or by its own effects as is demonstrated by Lascelles et al. (2010). 5 We demonstrated in this cross-over designed study in the healthy control cats a reflection of the supplemented fatty acids in the plasma fatty acid composition as is demonstrated by significant increased levels of eicosatrienoic acid (p=0.041), eicosatetraenoic acid (p=0.041), eicosapentaenoic acid (p=0.041), docosapentaenoic acid (p=0.049) and docosahexaenoic acid (p=0.041) after supplementation. Two client-owned cats disliked the taste of the test oil and/or the control oil. Furthermore two cats client-owned vomited occasionally while on oil supplementation, the data of these cats were not excluded because this occurred also before administration of the oils. In addition, one third of older cats (> 6 years of age) have a decreased capacity of fat digestion (Laflamme 2005). All these aspects may have contributed to a less then optimal uptake of the tested oils.

Zamprogno et al. (2010) showed that differences in clinical signs, found in their study, are indicative of effects on OA associated pain. This was concluded because the clinical signs were significantly different between cats with low radiographic scores and no signs of pain on manipulation, compared to cats with high radiographic scores and signs of pain on manipulation. We demonstrated significant differences in the same behavioral signs between cats on oil “A” and oil “B” as Zamprogno et al. (2010) demonstrated between cats with OA and healthy cats. Unfortunately, the groups in the study by Zamprogno (2010) were not age matched; the cats in the healthy group were younger, so age could also influence some parameters that are not necessarily related to OA, however, the study of Slingerland et al. (2011) demonstrated no correlation of OA-associated behavioral changes and age. Because the cats in our study had moderate OA, we could not calculate a correlation between OA score and grade of improvement on omega-3 fatty acid supplementation. The effects on OA associated pain in cats may be subjective because of owner interpretation, placebo effect, and better care effect, however owners and researchers were blinded and the study had a cross-over design, thus correcting for inter-observer differences. The improvement of behavior during play with other pets, jumping on objects, moments of grooming, and grooming time with the use of both oils may be attributed to placebo effect, and/or better care effect in analogy with the findings of Dobenecker et al. (2002) in dogs.

Slingerland et al. (2011) concluded that clinical examination of the larger peripheral joints had the highest sensitivity and specificity in relation to radiographically confirmed OA. As is known from previous studies (Clarke and Bennett 2006, Godfrey 2005, Lascelles

71 Part I

2010) there is no correlation between radiographic scores and clinical signs of OA in cats, also based on the fact that many cats resist being clinically investigated (Slingerland et al. 2011). The use of accelerometry to measure activity level as described by Lascelles et al. (2010) has several limitations and is not validated for use in cats. Force plate analysis has its limitations as well, because only the force of the limbs is measured and client- owned cats are difficult to be taught to walk on the force plate (Zumwalt et al. 2006). Radiographs in this study were used as an aid to include only cats with OA in this study; were not repeated as no meaningful changes are to be expected within the short period 5 of the study (Roush et al. 2010a).

The sample size was not determined by a power analysis before the start of the study. Because this study was a follow up of a previous study (Slingerland et al. 2011), from which we selected cats based on previous mentioned criteria, we knew precisely the presence and severity of OA in the joints of the appendicular skeleton in the investigated group of cats.

A questionnaire which has been used before to determine effects of OA on cat’s behavior and locomotion was used in this study and proved to be able to demonstrate subtle changes. The effects of the tested supplement, containing long-chain omega-3 fatty acids, such as a higher activity level (p=0.07), more walking up and down the stairs (p=0.07), less stiffness during gait (p=0.03), more interaction with the owner (p=0.07) and higher jumps (p=0.03), were visible in a randomized, double blinded, placebo controlled cross-over designed clinical trial.

Conclusion

A 10-week period on long-chain omega-3 fatty polyunsaturated fatty acids supplementation, (0.05g ETA, 1.53g EPA and 0.31g DHA per 1000kcal ME) changes the owner’s perception of some aspects of behavior and locomotion of cats with known naturally occurring osteoarthritis in comparison with a 10-week period on corn oil supplementation.

Acknowledgements

The authors thank Mr. S. van Huijzen and Dr. E. Teske for statistical analysis and interpretation, as well as Mrs. P. de Wit and Mrs. I.I.M. van Duiven for their technical assistance. The oils for this study were provided by Nutriceuticoils®, Belgium. The authors greatly acknowledge Luc Janssens for his contribution to the manuscript.

72 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

References • Dobenecker, B., Beetz, Y., Kienzle, E., 2002. A placebo-controlled double-blind study • Allan, G.S., 2000. Radiographic features of on the effect of nutraceuticals (chondroitin feline joint diseases. Veterinary Clinics of sulfate and mussel extract) in dogs with joint North America - Small Animal Practice 30, diseases as perceived by their owners. Journal 281-302. of Nutrition 132(6 Suppl 2),1690S-1691S. • Anderson, M., Beale, B., Boothe, D., Gaynor, J., • Freeman, L.M., 2010. Beneficial effects of McNamara, P., Sinclair, A., 2004. Roundtable omega-3 fatty acids in cardiovascular disease. on osteoarthritis in dogs and cats. Veterinary Journal of Small Animal Practice 51, 462-470. 5 Technician 25, 434-439. • Fritsch, D., Allen, T.A., Dodd, C.E., Jewell, • Beale, B.S., 2005. Orthopedic problems in D.E., Sixby, K.A., Leventhal, P.S., Hahn, K.A., geriatric dogs and cats. Veterinary Clinics of 2010. Dose-Titration effects of fish oil in North America - Small Animal Practice 35, Osteoarthritic dogs. Journal of Veterinary 655-674. Internal Medicine 24, 1020-1026. • Beale, B.S., 2004. Use of nutraceuticals and • Frost-Christensen, L.N., Lafeber, F., chondroprotectants in osteoarthritic dogs Mastbergen, S.C., Garssen, J., Hartog, and cats. Veterinary Clinics of North America - A., Hazewinkel, H.A.W., 2006. Effect of Small Animal Practice 34, 271-289. glycosamine combined with omega-3 • Blonk, M.C., Bilo, H.J., Nauta, J.J., Popp- fatty acids on the development of canine Snijders, C., Mulder, C., Donker, A.J., experimental osteoarthritis. Proceedings 1990. Dose-response effects of fish-oil 10th Congress of the European Society of supplementation in healthy volunteers. Veterinary and Comparative Nutrition 10, 155. American Journal of Clinical Nutrition 52, • Godfrey, D.R., 2005. Osteoarthritis in cats: A 120-127. retrospective radiological study. Journal of • Burns, K., 2006. Research targets conditions of Small Animal Practice 46, 425-429. older cats and dogs. Journal of the American • Gück, T., 2009. Therapeutic concepts in Veterinary Medical Association. 229, 482-483. osteoarthritis patients with antioxidants, • Cao, J., Schwichtenberg, K.A., Hanson, N.Q., chondroprotectiva and omega-3 fatty acids. Tsai, M.Y., 2006. Incorporation and clearance Kleintierpraxis 54, 703-717. of omega-3 fatty acids in erythrocyte • Gunew, M.N., Menrath, V.H., Marshall, membranes and plasma phospholipids. R.D., 2008. Long-term safety, efficacy and Clinical chemistry 52, 2265-2272. palatability of oral meloxicam at 0.01-0.03 • Clarke, S.P., Bennett, D., 2006. Feline mg/kg for treatment of osteoarthritic pain in osteoarthritis: A prospective study of 28 cases. cats. Journal of Feline Medicine and Surgery Journal of Small Animal Practice 47, 439-445. 10, 235-241. • Clarke, S.P., Mellor, D., Clements, D.N., • Hamilton, J.G, Comai, K., 1988. Rapid Gemmill, T., Farrell, M., Carmicheal, S., Bennett, separation of neutral lipids, free fatty acids D., 2005. Prevalence of radiographic signs and polar lipids using pre-packed silica Sep- of degenerative joint disease in a hospital Pak columns. Lipids 23, 1146-1149. population of cats. Veterinary Record 157, 793-799.

73 Part I

• Hansen, R.A., Ogilvie, G.K., Davenport, D.J., • Marshall, W.G., Hazewinkel, H.A.W., Mullen, Gross, K.L., Walton, J.A., Richardson, K.L., D., De Meyer, G., Baert, K., Carmichael, S., Mallinckrodt, C.H., Hand, M.S., Fettman, 2010. The effect of weight loss on lameness M.J., 1998. Duration of effects of dietary in obese dogs with osteoarthritis. Veterinary fish oil supplementation on serum Research Communications 34, 241-253. eicosapentaenoic acid and docosahexaenoic • Masson, S., Latini, R., Tacconi, M., Bernasconi, acid concentrations in dogs. American Journal R., 2007. Incorporation and washout of of Veterinary Research 59, 864-868. n-3 polyunsaturated fatty acids after diet 5 • Hardie, E.M., 1997. Management of supplementation in clinical studies. Journal of osteoarthritis in cats. Veterinary Clinics of Cardiovascular Medicine 8. North America - Small Animal Practice 27, • NRC, 2006. Nutrient Requirements of Dogs 945-953. and Cats. In: National Academy Press, • Hardie, E.M., Roe, S.C., Martin, F.R., 2002. Washington DC. Radiographic evidence of degenerative joint • Retra, K., Bleijerveld, O.B., Gestel, R.A. van, disease in geriatric cats: 100 cases (1994– Tielens, A.G.M., Hellemond, J.J. van, Brouwers, 1997). Journal of the American Veterinary F.F., 2008. A simple and universal method Medical Association 220, 628–632. for the separation and identification of • Kates, M., 1986. Techniques of Lipidology, phospholipid molecular species. Rapid Elsevier, Philadelphia, 125. Communications in Mass Spectrometry 12, • Kerwin, S.C., 2010. Osteoarthritis in cats. 1853-62. Topics in Companion Animal Medicine 25, • Rothschild, B.M., Rothschild, C., Woods, R.J., 218-223. 1998. Inflammatory arthritis in large cats: An • Laflamme, D.P., 2005. Nutrition for aging expanded spectrum of spondyloarthropathy. cats and dogs and the importance of body Journal of Zoo and Wildlife Medicine 29, condition. Veterinary Clinics of North America 279-284. - Small Animal Practice 35, 713-742. • Roush, J.K., Cross, A.R., Renberg, W.C., Dodd, • Lascelles, B.D.X., 2010. Feline degenerative C.E., Sixby, K.A., Fritsch, D.A., Allen, T.A., Jewell, joint disease. Veterinary Surgery 39, 2-13. D.E., Richardson, D.C., Leventhal, P.S., Hahn, • Lascelles, B.D.X., 2008. Incidence of feline DJD K.A., 2010a. Evaluation of the effects of dietary and what the radiographic changes mean. supplementation with fish oil omega-3 In: 14th Congress of the European Society of fatty acids on weight bearing in dogs with Veterinary Orthopedics and Traumatology, osteoarthritis. Journal of the American Munich, pp. 126-127. Veterinary Medical Association 236, 67-73. • Lascelles, B.D.X., DePuy, V., Thomson, A., • Roush, J.K., Dodd, C.E., Fritsch, D.A., Allen, T.A., Hansen, B., Marcellin-Little, D.J., Biourge, V., Jewell, D.E., Schoenherr, W.D., Richardson, Bauer, J.E., 2010. Evaluation of a therapeutic D.C., Leventhal, P.S., Hahn, K.A., 2010b. diet for feline degenerative joint disease. Multicenter veterinary practice assessment Journal of Veterinary Internal Medicine 24, of the effects of omega-3 fatty acids on 487-495. osteoarthritis in dogs. Journal of the American Veterinary Medical Association 236, 59-65. • Rychel, J.K., 2010. Diagnosis and treatment of osteoarthritis. Topics in Companion Animal Medicine 25, 20-25.

74 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

• Scarlett, J.M., Donoghue, S., 1998. • Zumwalt, A.C., Hamrick, M., Schmitt, D., 2006. Associations between body condition and Force plate for measuring the ground reaction disease in cats. Journal of the American forces in small animal locomotion. Journal of Veterinary Medical Association 212, 1725- Biomechanics 39, 2877-2881. 1731. • Simopoulou, T., Malizos, K.N., Iliopoulos, D., Stefanou, N., Papatheodorou, L., Ioannou, M., Tsezou, A., 2007. Differential expression of leptin and leptin’s receptor isoform (Ob-Rb) 5 mRNA between advanced and minimally affected OA cartilage; effect on cartilage metabolism. Osteoarthritis Cartilage 15, 872-83. • Slingerland, L.I., Hazewinkel, H.A.W., Meij, B.P., Picavet, P., Voorhout, G., 2011. Cross-sectional study of the prevalence and clinical features of osteoarthritis in 100 cats. The Veterinary Journal 187, 305-309. • Taylor, P.M., Robertson, S.A., 2004. Pain management in cats - Past, present and future. Part 1. The cat is unique. Journal of Feline Medicine and Surgery 6, 313-320. • Trepanier, L., 2009. Idiopathic inflammatory bowel disease in cats. Rational treatment selection. Journal of Feline Medicine and Surgery 11, 32-38. • Zamprogno, H., Hansen, B.D., Bondell, H.D., Sumrell, A.T., Simpson, W., Robertson, I.D., Brown, J., Pease, A.P., Roe, S.C., Hardie, E.M., Wheeler, S.J., Lascelles, D.X., 2010. Item generation and design testing of a questionnaire to assess degenerative joint disease-associated pain in cats. American Journal of Veterinary Research 71, 1417-1424. • Zainal, Z., Longman, A.J., Hurst, S., Duggan, K., Caterson B., Hughes, C.E., Harwood, J.L., 2009. Relative efficacies of omega-3 polyunsaturated fatty acids in reducing expression of key proteins in a model system for studying osteoarthritis. Osteoarthritis Cartilage 17, 896-905.

75 Part I

Appendix 1 Questionnaire

1. Is your cat an indoor cat, an outdoor cat, or both? 2. How long is your cat outside on average: 0 hour, 0-3 hours, more then 3 hours? 3. Did you notice differences in time that you are allowed to groom your cat? (More, a bit more, equal, a bit less, less) 4. Does your cat urinate or poop outside the litter box? (Yes/no) 5. Did you notice a change in usage of the litter box? (More, a bit more, equal, a bit less, 5 less) 6. Did you notice behavioral changes during play with other pets? (More, a bit more, equal, a bit less, less) 7. Did you notice a change in activity level? (More, a bit more, equal, a bit less, less) 8. Did you notice differences in jumping on the table/couch/windowsill etc.? (More, a bit more, equal, a bit less, less) 9. Is there a change in jumping height? (More, a bit more, equal, a bit less, less) 10. Are the spots that the cat likes best easy accessible for the cat? (More, a bit more, equal, a bit less, less) 11. Did you notice a change in walking up the stairs? (More, a bit more, equal, a bit less, less) 12. Did you notice a change in walking down the stairs? (More, a bit more, equal, a bit less, less) 13. Do you notice the cat being stiffer during gait? (More, a bit more, equal, a bit less, less) 14. Did you notice a change in aggressive behavior? (More, a bit more, equal, a bit less, less) 15. Did you notice a change in satisfaction of your cat? (More, a bit more, equal, a bit less, less) 16. Did you notice a change in greeting people? (More, a bit more, equal, a bit less, less) 17. Did you notice a change in the time the cat wants to be petted? (More, a bit more, equal, a bit less, less) 18. Did you notice a change in the interaction with other cats? (More, a bit more, equal, a bit less, less) 19. Did you notice a change in the number of moments on which the cat is grooming itself? (More, a bit more, equal, a bit less, less) 20. Did you notice a change in grooming time? (More, a bit more, equal, a bit less, less) 21. Has your cat ever been hit by a car? (Yes/no) 22. Has the cat been lame in the past? (Yes/no) 23. Does your cat show lameness at the moment? (Yes/no)

Questions 1, 2, 21 en 22 were asked once.

76 The effect of dietary long-chain omega-3 fatty acid supplementation on owner’s perception of behaviour and locomotion in cats with naturally occurring osteoarthritis

77 Part II: Vitamin A and D requirements in relation to skeletal health

78 Chapter 6

The interaction of vitamin A and vitamin D with emphasis on bone metabolism

79 Part II

80 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

Historical facts

Both vitamin A and vitamin D were discovered in the early 20th century (Wolf 2004, Semba 2012), however, the symptoms of their deficiency were described earlier. In man, rickets (vitamin D deficiency) was already described by F. Glisson in 1650 (Wolf 2004), and corneal ulcers and high mortality (vitamin A deficiency) was described by F. Magendie in 1816 (Semba 2012). Treatment of rickets with cod-liver oil (being a dietary source of vitamin D) has been already described by D. Scheutte in 1824 (Wolf 2004), but the existence of “unexpected essential dietary factors” was postulated first by Hopkins (1906). Treatment of xerophthalmia in young rats with a “fat-soluble, non-saponifiable factor from butterfat” was described by McCollum and Davis (1914) and this factor was 6 called by them fat-soluble factor A, later vitamin A. The existence of fat-soluble factor A, together with the known pathology of rickets formed the basis of the experiment of Mellanby (1919), who fed puppies a diet of low-fat milk and bread. Skeletal pathology was diagnosed in these puppies, which was similar to human rickets. Mellanby also added yeast (B-vitamins) and orange juice (vitamin C) to the diet of these puppies, but these additives did not prevent development of rickets. Addition of butterfat or cod-liver oil were effective in prevention of rickets, so Mellanby concluded that rickets was the result of a dietary deficiency of anti-rachitic factor, which is either fat-soluble factor A, or a substance with a similar distribution (Mellanby 1919). Heat and oxidation induced inactivation of vitamin A was discovered by Hopkins in 1920, as he demonstrated that baked butterfat did not prevent xerophthalmia and mortality in rats. This experiment was followed by McCollum et al. in 1922 demonstrating that heated, oxidized cod-liver oil did not prevent xerophthalmia, but cured rickets in rats. Based on this experiment it was concluded that fat-soluble factor A consisted of 2 different substances. Because B-vitamins and vitamin C were already identified, the anti-rachitic factor was named vitamin D (Wolf 2004, Semba 2012).

Vitamin A metabolism

All dietary vitamin A originates from carotenoids synthesized by plants. The conversion of carotenoids to vitamin A requires oxidative cleavage of the carotenoid molecule by the enzyme β-carotene 15,15’-dioxygenase, which produces 2 molecules of retinol. Consumption of prey, especially its liver, leads to consumption of pre-formed vitamin A. Vitamin A is ingested as an ester to be hydrolyzed by pancreatic enzymes in the intestine. Vitamin A is absorbed from the gut by protein mediated and passive diffusion. Free retinol is sequestered by cellular retinol-binding proteins (RBPs) in the enterocyt. Retinoids may be esterified to retinyl esters or oxidized to retinal. Retinyl esters and unesterified retinol are transported into chylomicrons. After being stripped of their triacylglycerols by lipoprotein lipase, the remnants are taken up by endocytosis into hepatocytes. The main metabolic pathways of vitamin A take place in liver and kidneys. In the liver different pathways are being followed by retinyl esters and unesterified retinol: 1) binding to RBPs and transportation in the bloodstream, 2) re-esterification, 3) storage as retinyl esters in

81 Part II

lipid droplets in stellate cells 4) further metabolization. In the kidney, retinol and retinyl esters are excreted in the urine.

Comparative aspects of vitamin A metabolism

Cats and ferrets are unable to efficiently produce vitamin A from carotenoids in their livers, so their diet must contain pre-formed vitamin A (Morris 2002) which can be found in several feed stuff (Table 1). Vitamin A in cats is excreted as retinylesters in the urine in much smaller amounts compared to other species and cats have higher vitamin A storage (as retinyl esters) in the kidneys, compared to other species (Raila and Schweigert 2002) to 6 protect them from hypovitaminosis A. Carnivores, like other species, prevent high plasma levels of retinol as a consequence of the consumption of large amounts of vitamin A from their prey by converting retinol to retinyl esters. Circulating retinyl esters are indicative of hypervitaminosis A in most species, but are considered normal in carnivorous species (Raila et al. 2001), with the exception of the family Hyenidae and suborder Pinnipedia that only have retinol in their plasma (Raila et al. 2001). High physiological circulating levels of retinyl esters were also demonstrated in arctic predators (e.g. polar bears, arctic foxes, bearded seals, ringed seals, glaucous gulls, fulmars, guillemots, puffins, reindeers and ptarmigans) that consume prey with 10-20 fold higher vitamin A concentrations compared to prey from other areas (Senoo et al. 2012).

Effects of vitamin A on tissues Vision Vitamin A is involved in the visual cycle, as 11-cis retinal is attached to rhodopsin, and forms metarhodopsin II under the influence of light, which then reacts with transducin, leading to a reduction of cyclic guanosine monophosphate. Consequentially signal transduction through the optical nerve takes place. In this visual process, vitamin A is efficiently recycled, so this does not contribute much to the daily vitamin A requirements.

Cellular differentiation, morphogenesis, and growth Historically, xerophthalmia is one of the first described symptoms of vitamin A deficiency (Semba 2012). Later on, hyperkeratosis of the mucus secreting cells of the respiratory tract, gastrointestinal tract and reproductive tract were described to be caused by vitamin A deficiency. The cellular differentiation is mediated by retinoids through the oxidation of all-trans retinol associated with RBPs to all-trans retinoic acid (ATRA). ATRA is transported to the nucleus where it is bound to the retinoic acid receptor (RAR) or retinoid X receptor (RXR) (Morris-Kay et al. 1994). RXR can form heterodimers with vitamin D receptor (VDR), thereby influencing osteoblasts. The binding of ATRA to RXR-VDR dimers can only occur if VDR is charged with 1,25-dihydroxyvitamin D. RXR can also form heterodimers with thyroid receptor (TR). The binding of ATRA to RXR-TR dimers can only occur if TR is charged with triiodothyronine, thereby influencing growth and morphogenesis. After binding to the dimers, several response elements are activated, influencing cellular differentiation, morphogenesis and growth.

82 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

Table 1: Vitamin A content of feed stuff in IU per g as fed Cod liver oil 1,000 Liver - Turkey 269 - Beef, pork, fish 217 - Chicken 110 Dandelion greens 56 Sweet potato 32 Carrot 28 6 Broccoli leaf 27 Butter 23 Kale 23 Spinach 16 Pumpkin 13 Collard greens 10 Cheddar cheese 9 Cantaloupe melon 6 Egg 5 Apricot 3.2 Papaya 1.8 Mango 1.3 Pea 1.3 Broccoli florets 1.0 Milk 0.9

Immune response High mortality was described in rats that were fed a vitamin A deficient diet. The effects of vitamin A on immunity have also been demonstrated in rats (Ross and Hammerling 1994). Vitamin A deficiency was associated with attenuated B-cell and T-cell responses to human rotavirus in a gnotobiotic piglet model (Chattha et al. 2013). The anti-viral effects of vitamin A supplementation was also demonstrated in ferrets to canine distemper virus (Rodeheffer et al. 2007).

Effects of vitamin A on bone

Vitamin A plays an important role in the shaping of the bone by promoting osteogenic differentiation, and cartilage maturation and mineralization. Vitamin A also increases matrixmetalloproteinase-13 (MMP-13) and aggrecanase activity in joint cartilage as

83 Part II

well as induced mineralization, and up-regulation of osteonectin and osteopontin (both structural non-collagenous proteins), produced by osteoblasts. Furthermore, ATRA alters chondrocyte gene expression of type I, II, IX, and XI collagen towards a non-chondrogenic and osteoarthritis-like phenotype (Davies et al. 2009).

In 5-weeks old rats that were fed high dietary levels of vitamin A (i.e. 513,600 μg per kg diet for 8 days), increased endosteal mineralization and osteoblast numbers in diaphyseal bone were noted, as well as increased osteomodulin (Omd) expression (bone-specific extracellular matrix protein), and Wnt signaling in osteoblasts from the endosteum, but reduced Wnt signaling in osteoblasts from the periosteum. Reduced cortical and increased 6 endosteal bone formation were demonstrated, which caused increased mineralization and stiffness of the endosteal bone, and thinner cortices, without differences in total bone mineral density (Lind et al. 2012). Inhibition of collagen synthesis, as reported by Hough et al (1988) in growing rats consuming 10,000 and 25,000 IU vitamin A per day, respectively, may weaken insertion sites of muscles and tendons making the bone more prone to trauma, even during normal movements. Possibly this weakening of insertion sites is due to the up-regulation of MMP-9 activity, a gelatinase enzyme, by vitamin A (Lackey and Hoag 2010).

Clinical signs of vitamin A deficiency and excess Vitamin A deficiency Vitamin A deficiency is caused by a diet that is deficient in vitamin A and/or carotenoids, and/or by inactivation of vitamin A due to heating and/or oxidation. The effects of vitamin A deficiency are mostly seen in cells with a high turnover rate, like cornea, respiratory, and gastrointestinal epithelial cells. Signs of vitamin A deficiency include anorexia, weight loss, ataxia, xerophthalmia, conjunctivitis, corneal opacity and ulceration, skin lesions (i.e. hyperkeratosis), metaplasia of the bronchiolar epithelium, pneumonitis, and increased susceptibility to infections, as well as abnormal bone growth and tooth development (Steenbock et al. 1921, Stimson and Hedley 1933, Frohring 1935a,b, 1937, Crimm and Short 1937, Russell and Morris 1939, Gershoff et al. 1957, Singh et al. 1965).

Comparative aspects of Vitamin A deficiency In kittens, cleft palates were noted when queens were fed diets containing 600-1200 μg retinol per kg on dry matter basis (d.m.b.). A diet containing 1800 μg retinol per kg d.m.b. (400 μg per 1000kcal ME) prevents development of these deformities in kittens (Morris, personal communication). In young growing dogs, deficiency of vitamin A resulted in defective remodeling of the cranial foramina, causing stenosis of the cochlear nerve, accompanied by bony growth in the modiolus, and overgrowth of the internal periosteal layer of the capsule of the organ of Corti (Mellanby, 1938).

Vitamin A excess Several species demonstrate skeletal deformities due to hypervitaminosis A, especially in neonates and growing animals.

84 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

Table 2: Current vitamin A requirements of dogs and cats Table 3: Vitamin D content of feed stuff in IU per mL in oil and in per g as fed for other feed stuff Minimum requirements (in IU per 1000kcal ME): Dog Cod liver oil 90.6 - Adult 1,515 Catfish (wild) 5 - Early growth (<14 weeks) 1,250 Sardines, canned in oil, drained 5 - Late growth (≥14 weeks) 1,250 Salmon, cooked 3.6 Cat Mackerel, cooked 3.45 - Adult 833 Tuna, canned in oil 2.35 Eel, cooked 2.00 - Growth/reproduction 2,250 6 Whole egg 0.33 Maximum allowance (in IU per 1000kcal ME): Beef liver, cooked 0.15 Dog 100,000 Cat 100,000 - Adult/growth 100,000 - Reproduction 83,325

Table 4: Current vitamin D requirements of dogs and cats

Minimum requirements (in IU per 1000kcal ME): Dog - Adult 138 - Early growth (<14 weeks) 138 - Late growth (≥14 weeks) 125 Cat - Adult 62.5 - Growth/reproduction 188

Maximum allowance (in IU per 1000kcal ME): Dog 800 Cat 7,500

85 Part II

Neonates High vitamin A intake during pregnancy is associated with cleft palate in newborn rodents (Nanda 1970), humans (Rothman et al. 1995), pigs (Wolke et al. 1968), dogs (Wiersig et al. 1967), and cats (Freytag et al. 2003). Immatures In juvenile gilthead sea bream, high dietary intake of vitamin A (i.e. 780,000 μg retinol per kg diet for 3 months) caused degradation of collagen in the growth zones of vertebral end plates and the trabecular bone layers, possibly due to increased collagenase synthesis and collagen degradation in bone by the actions of retinoic acid (Fernandez et al. 2012). When the more brittle bone tissue of the vertebrae is stressed (e.g. in case of lordosis), 6 this leads to the appearance of chondroid bone, which meets the demand for accelerated growth to establish shear-resistance (Huysseune 2000). Hypervitaminosis A alters bone homeostasis through an accelerated bone mineralization, which might increase mechanical load between adjacent vertebrae and induce vertebral compression/fusion through a bone remodeling process (Fernandez et al. 2012).

In juvenile lambs, high vitamin A intake (i.e. 450,000 μg of retinol per kg milk replacer product, fed for 2 weeks) resulted in early or irregular physeal closure of long bones and vertebrae (Reichel et al. 2012). In calves and young growing dogs high vitamin A intake can cause the premature closure of growth plates due to suppression of chondrocyte differentiation and proliferation (Cho et al. 1975, Yamamoto et al. 2003).

Vitamin A excess in cats Neonates In kittens, cleft palates were noted when queens were fed 306,000 μg retinol per kg diet d.m.b. (Freytag 2001, Freytag et al. 2003). Immatures Feeding a diet largely or totally composed of raw liver to growing kittens from weaning to adulthood leads to extensive osteocartilagenous hyperplasia of the cervical vertebrae (Seawright et al. 1967, 1970). Shorter bones with damaged epiphyseal plates and osteoporosis were noted (Clark et al. 1970, Clark 1973), as well as tooth loss due to retarded development of the alveolar processes (Seawright and Hrdlicka 1974). Adults In adult cats, hypervitaminosis A is characterized by typical bone lesions (i.e. hyperostosis of the cervical vertebrae). The prevalence of hypervitaminosis A in cats is unknown, and there is no breed or sex predilection. Hypervitaminosis A in cats is caused by chronic consumption of raw liver rich in vitamin A (Polizopoulou et al. 2005) or by vitamin A supplements (Hazewinkel 1994). Clinical signs are lameness, sometimes followed by paresis and paralysis, an unkempt hair coat, and sometimes anisocoria and/or Horner’s syndrome (Polizopoulou et al. 2005). In limbs, bone lesions originate from the insertion areas, resulting in restricted movement and finally extra-articular new bone formation leading to ankylosis. On radiographs deforming cervical spondylosis and formation of osteophytes and exostoses at the insertion site of tendons, ligaments, and joint

86 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

capsules can be seen. Hypervitaminosis A is therefore considered as a secondary form of osteoarthritis in the vertebrae and the large joints (Lascelles 2010). A case series of 8 cats with hypervitaminosis A revealed that the vertebrae, the elbow joints, and the stifle joints are mostly affected (Hazewinkel 1994). Histopathologically the bony lesions have a subperiostal origin with apposition of new woven bone and loss of the cortical bone pattern (Franch et al. 2000, Polizopoulou et al. 2005). Earliest observed changes in vertebrae consist of cartilaginous and osseous hyperplasia not associated with inflammation, finally resulting in periarticular new bone formation. Despite the bony lesions, the articular space is preserved (Franch et al. 2000). Feeding a low vitamin A diet, and pain management with non-steroidal anti-inflammatory drugs can be used as treatment in cats with chronic hypervitaminosis A, which results in clinical improvement, 6 however the bony changes are irreversible (Hazewinkel 1994).

Vitamin A excess in dogs Immatures Growing dogs seem to be more sensitive to high vitamin A intake, compared to adult dogs. High vitamin A intake (i.e. 6 days per week 300,000 IU of vitamin A per kg body weight for 30 days) in young dogs was studied in 2 month-old greyhound puppies. Radiography of the long bones revealed marked narrowing of the physeal cartilage and reduction in the mineral density and thinning of the cortices. Histology revealed a disturbance in endochondral ossification (Maddock et al. 1949). Another study on high vitamin A intake was done in Labrador retriever puppies. Different treatment groups were studied, but only in the group receiving 30,000 μg retinol per kg body weight once a week parenterally for 10 weeks, then oral 90,000 μg retinol per kg body weight per day during 88 days, demonstrated bone pathology. Long bones were decreased in length and thickness because of premature closure of the physeal plate. Osteoporosis of the diaphysis was also present. Interestingly, growing dogs that were administered vitamin A-D-E combinations (30,000 to 60,000 μg retinol + 3750 mg vitamin D2 + 100 IU vitamin E per kg body weight one a week parenterally for 14 weeks) were less severely affected compared to dogs that were fed solely vitamin A (Cho et al., 1975). A recent safety evaluation was performed in puppies from 8 weeks of age that were fed a diet with 100,000 IU of vitamin A per 1000kcal ME for 52 weeks and reported no adverse effects on blood parameters and collagen I by dual energy x-ray absorptiometry (Morris et al. 2012). Adults Adult dogs are considered to be quite resistant to high vitamin A intake. In adult dogs that were fed a diet with 270,000 μg retinol per kg diet d.m.b. for 1 year, no adverse changes in the tibia mineralization, as measured by computer tomography, were noted (Cline et al. 1997).

Vitamin D metabolism

Dietary vitamin D originates from plants (vitamin D2), prey (vitamin D3), or is synthesized in the skin under the influence of ultraviolet B light (UVB). Synthesized vitamin D is

87 Part II

bound to vitamin D binding proteins and transported to target organs. Dietary vitamin D is absorbed from the gut by protein mediated and passive diffusion. It is transported to the liver in chylomicrons (NRC 2006). Vitamin D is firstly metabolized in the liver by 25-hydroxylase, which is weakly regulated, and therefore, 25-hydroxy vitamin D (25-OHD) reflects dietary vitamin D intake (How et al. 1994, Morris 1999). The 25-OHD is either transported to the kidney for further metabolism facilitated by 1-alpha-hydroxylase to the most active metabolite 1,25-dihydroxy vitamin D (1,25DHCC), or is metabolized by 24-hydroxylase is several tissues to 24,25-dihydroxy vitamin D (24,25DHCC), or it can be stored in the liver (Hazewinkel and Tryfonidou 2002). The hydroxylation of 25-OHD is regulated on endocrine level by plasma Ca and P concentrations, as well as by other 6 calciotropic hormones (including parathyroid hormone, calcitonin, growth hormone, insulin-like growth factor 1, and fibroblast growth factor 23). (Tryfonidou et al. 2003b). The hydroxylation of 1,25DHCC to 1,24,25-trihydroxy-vitamin D (1,24,25THCC), and of 25- OHD to 24,25DHCC by 24-hydroxylase takes place in different tissues (including, bone, liver, and kidney). 1,24,25THCC is excreted in the urine.

Comparative aspects of vitamin D metabolism

In contrast to most herbivorous and omnivorous species, cats and dogs are unable to synthesize adequate amounts of vitamin D under ultraviolet B light (How et al. 1994). Morris (1999) demonstrated that cats are able to synthesize vitamin D after administration of a delta-7-dehydrocholesterol reductase inhibitor. Morris (1999) concluded that strict carnivores do not need to convert 7-dehydrocholesterol (7-DHC) into vitamin D, as they already consume large amounts of vitamin D when eating whole prey. In stead, carnivores use 7-DHC for other purposes, like cholesterol synthesis.

Effects of vitamin D on tissues

The main functions of vitamin D are related to calcium (Ca) and phosphorus (P) homeostasis, and bone metabolism. It is also required by the immune system.

Cardiovascular system In human medicine, low 25OHD plasma levels were associated with left ventricular hypertrophy vascular dysfunction, and renin-angiotensin system activation, however the causal relationship remains to be established (Al Mheid et al. 2013).

Central nervous system In humans, an association of vitamin D deficiency and depression has been demonstrated due to hyperparathyroidism as a consequence of low vitamin D intake and low exposure to UVB (Hoogendijk et al. 2008). The mood is usually improved after treatment of hyperparathyroidism (Watson et al. 2002). Furthermore, both low and high plasma concentrations of vitamin D were associated with increased risk of schizophrenia in neonatal children, probably due to low 1,25DHCC activity in the human brain (McGrath

88 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

et al. 2010). In rodent models transient prenatal vitamin D deficiency results in persistent changes in adult brain structure, neurochemistry, and behavior (Féron et al. 2005).

Immune system Diverse effects of vitamin D on cells of innate and adaptive immunity have been reported including pro-allergic T helper type 2 (Th2) polarization and cytokine production (IL-4, IL- 5, and IL-10), but conversely also anti-allergic effects such as inhibition of IgE production by B cells, inhibition of dendritic cell (DC) maturation and migration, or conversion of CD4+ T cells to T regulatory cells (Von Guten et al. 2013, Benson et al. 2012). Elevated IgE production has been observed in humans with both extremely low or high vitamin D concentrations, suggesting that unbalanced vitamin D levels deviating from a certain 6 threshold could lead to aberrant immune responses (Hyppönen et al. 2009).

Effects of vitamin D on bone

24,25DHCC plays an important role in activating resting chondrocytes towards migration, and makes them responsive for 1,25DHCC, which is responsible for the terminal chondrocyte differentiation, and cartilage matrix production and mineralization (Schwarz et al. 1995). In addition, 24,25DHCC acts on periosteal osteoblasts, increasing periosteal new bone formation, but without increased osteoclast activity due to suppression of parathyroid hormone (PTH) (Van Leeuwen et al. 2001).

1,25DHCC affects bone by increasing the resorption of mineralized osteoid to release Ca and P and by increasing the mineralization of newly formed osteoid and cartilage with Ca and P (Schwarz et al. 1995). Receptors to 1,25DHCC are found in osteoblasts, but not in osteoclasts (Imai et al. 2013). The effect of 1,25DHCC on osteoblasts results in change in shape, which allows osteoclasts to make contact to the bone and resorb it, thus releasing Ca and P from the bone and increase plasma Ca and P levels (Tryfonidou et al. 2010). Furthermore 1,25DHCC regulates osteoblast proliferation and activity (Narbaitz et al. 1983, Norman et al 2002). 1,25DHCC is also responsible for increased active Ca and P absorption from the gut by activating alkaline phosphatase, calbindin, and the plasma membrane Ca pump (Christakos et al. 1996), and increased Ca and P reabsorption from the pre-urine (Winaver et al. 1980, Yanagawa and Lee 1992).

Clinical signs of vitamin D deficiency and excess Vitamin D deficiency in cats Vitamin D deficiency in cats causes decreased growth, enlarged growth plates, osteopenia and fibrous osteodystrophy. Due to poor mineralization of newly formed osteoid, the limbs bend, or spontaneous pathological fractures may occur. These signs can be prevented by a dietary concentration of 3.125 μg vitamin D per kg diet (Morris et al. 1999).

89 Part II

Vitamin D deficiency in dogs In dogs with rickets, the physeal plates of the bones are enlarged and long bones are bended. There is osteopenia and the physis of the distal ulna increases in width. Diagnosis rickets could be confirmed by low plasma 25-vitamin D levels (How et al. 1994). Studies in Great Dane puppies by Tryfonidou et al. 2002 demonstrated no clinical effects when feeding a diet containing 12.5 μg of vitamin D per kg diet, which confirmed the minimum requirements of NRC 1985.

Vitamin D excess in cats Vitamin D intoxication in cats has been reported as a result of formulation errors, 6 consumption of fish viscera, or ingestion of rodenticides. Hypercalcemia, renal failure, and soft tissue calcifications have been reported (Sato et al. 1993, Haruna et al. 1992, Moore et al. 1988), as well as calcium oxalate uroliths (Morita et al. 1995). A safety study, that has set the NRC (2006) maximum allowance for vitamin D in cats at 750 μg of vitamin D per kg diet d.m.b., demonstrated no effects on a diet containing 846 μg of vitamin D per kg diet d.m.b. when it was fed to 57 cats (17 adults and 40 kittens) for a 2 year period (Sih et al. 2001).

Vitamin D excess in dogs In a study with Great Dane puppies, a diet containing 1,350 μg of vitamin D per kg resulted in irregular growth plates and severely disturbed endochondral ossification (Tryfonidou et al. 2003a). Puppies given a diet containing 100 μg cholecalciferol per kg diet also demonstrated impaired endochondral ossification (Tryfonidou et al. 2002). Both studies did not demonstrate an increase in plasma Ca and P levels, nor soft tissue calcifications, and the symptoms were less severe in the study using the lower dosage of vitamin D.

Interactions in vitamin A and vitamin D metabolism

Some studies demonstrated delayed/reduced pathology when vitamin A and vitamin D were both supplemented (Moore and Wang 1945) or when vitamin A, vitamin D, and vitamin E were supplemented together, compared to supplementation of vitamin A only (Cho et al. 1975). On the contrary, the skeletal hyperostosis seen in cats consuming raw liver and/or vitamin A and vitamin D containing supplements were not demonstrated in queens that were fed high levels of retinol (i.e. 306,000 μg retinol per kg diet) for a 2-3 year period. This might indicate that the skeletal hyperostoses seen in cats are not due to high vitamin A intake alone, but only in combination with increased vitamin D intake.

Absorption Both vitamin A and vitamin D are absorbed from the gut by diffusion, and transported through lymph vessels to the liver where they are metabolized and stored, or released. Interaction between vitamin A and vitamin D on absorption level (diffusion) can be expected, however, to date, this interaction has not been studied.

90 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

Carotenoids 7DHC β-carotene 15,15’-dioxygenase UVB+temperature Production Retinol Vitamin D Retinol dehydrogenase 25-hydroxylase Activation Retinaldehyde 250HD Retinaldehyde dehydrogenase CYP 24 1-α hydroxylase

ATRA 24,25DHCC 1,25DHCC 6 CYP 26A1 CYP 24 CYP 26B1 Inactivation Hydroxylated RA 1,24,25THCC

ATRA 1,25DHCC RAR RXR Nuclear receptors VDR RXR RARES VDREs

Fig. 1: Metabolic pathways of vitamin A and D 7DHC = 7-dehydrocholesterol; UVB = ultraviolet B light; 25OHD = 25-hydroxyvitamin D; CYP = cytochrome P enzyme; ATRA = all-trans retinoic acid; 24,25DHCC = 24,25-dihydroxyvitamin D; 1,25DHCC = 1,25-dihydroxyvitamin D; RA = retinoic acid; 1,24,25THCC = 1,24,25-trihydroxy vitamin D; RAR = retinoic acid receptor; RXR = retinoid X receptor; VDR = vitamin D receptor; RAREs = retinoic acid response elements; VDREs = vitamin D response elements

Vitamin A can be formed out of carotenoids or is ingested as either retinol or as retinyl esters from the diet. Retinol needs to be activated to all-trans retinoic acid (ATRA) to become effective. ATRA can be bound to retinoic acid receptor to stimulate target genes (retinoic acid response elements). Inactivation of ATRA takes place by enzymatic hydroxylation. Vitamin D can be formed out of 7- dehydrocholesterol or is ingested from the diet. In the liver, vitamin D is coinverted to 25-hydroxyvitamin D (25OHD). 25OHD needs to be activated to 1,25-dihydroxyvitamin D (1,25DHCC) to become effective. 1,25DHCC can be bound to vitamin D receptor to stimulate target genes (vitamin D response elements). Inactivation of vitamin D takes place by enzymatic hydroxylation

91 Part II

Cytochrome-P450 Both vitamin A and vitamin D are metabolized by enzymes of the Cytochrome-P450 (CYP450) super family, which catalyzes oxidation and hydroxylation reactions on lipophilic substrates to get hydrophilic substrates to produce active metabolites and/or to improve renal elimination of the substrate from the body. CYP2C8 is involved in both the activation of retinol and the inactivation of ATRA, whereas CYP2C9 is only involved in the inactivation of ATRA (Marill et al. 2003). CYP2R1 is involved in the 25-hydroxylase of vitamin D. High vitamin A levels result in an up-regulation of CYP24, the CYP450 enzyme that metabolizes 25OHD into the 24,25DHCC metabolite (Allegretto et al. 1995). This might be indirectly by up-regulation of fibroblast growth factor 23 (FGF23) by 6 ATRA. FGF23 also stimulates CYP24 and inhibits CYP27B1 (1-alpha-hydroxylase) (Quarles 2012). Up-regulation of FGF23 by vitamin A has not been investigated, but it is known that FGF23 is up-regulated by VDR-RXR heterodimer when 1,25DHCC is bound as ligand. Chronic excessive vitamin D intake also increases CYP24, and thus increases 24,25DHCC production.

CYP26A1 and CYP26B1 in osteoblasts are involved in bone metabolism and are regulated by vitamin A (Spoorendonk et al. 2008). CYP26A1 down-regulates the intracellular ATRA concentration, which is able to influence gene expression mediated by RAR and RXR. In case of hypervitaminosis A the regulation of intracellular ATRA cannot be controlled anymore, leading to overexpression of RAR and RXR response elements, resulting in teratogenic effects of hypervitaminosis A in mice embryos (Abu-Abed et al. 2001). CYP26B1 regulates the intracellular inactivation of ATRA. In zebra fish CYP26B1 gene mutation allows ATRA to access osteoblasts, and these fish develop severe ankylosing spondylosis (Spoorendonk et al. 2008). In rat, a direct effect of vitamin A on osteoclast activity is proposed, resulting in osteoporosis (Kneissel et al. 2005). Vitamin A is known to up- and down-regulate the expression of CYP genes by the binding of RXR or RAR to response elements on different target genes.

So far, the only known interaction between vitamin A and vitamin D on CYP enzymes are the up-regulation of CYP24 by vitamin A, which stimulates hydroxylation of 25OHD and 1,25DHCC and increased 24,25-DHCC production and thus increased (periosteal) new bone formation.

Receptors RXR can form heterodimers with vitamin D receptor (VDR), but only when VDR is charged with 1,25DHCC (Olson 2001). These heterodimers are able to bind to specific response elements for VDR and RXR on several genes. The effects of excess vitamin A on bone might be the result of the same mechanisms, as RXR and VDR are both present in chondrocytes and osteoblasts (Imai et al. 2013, Pike et al. 2012). RXR can be activated by ATRA, which suppresses bone resorption in the presence of 1,25DHCC, possibly due to inhibition of receptor-activator nuclear factor Κβ ligand (RANKL), positive Wnt signaling, and promoting differentiation of mesenchymal stem cells into primary osteoblasts, as

92 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

ATRA FGF23 1,25DHCC 250HD

1-α hydroxylase

CYP 24

1,24,25THCC 24,25DHCC

Fig. 2: Interactions between vitamin A and vitamin D ATRA = all-trans retinoic acid; 25OHD = 25-hydroxyvitamin D; 1,25DHCC = 1,25-dihydroxyvitamin D; FGF = fibroblast growth factor; 6 CYP = cytochrome P enzyme; 24,25DHCC = 24,25-dihydroxyvitamin D; 1,24,25THCC = 1,24,25-trihydroxy vitamin D

All-trans retinoic acid (ATRA) and 1,25-dihydroxyvitamin D (1,25DHCC) both stimulate the formation of fibroblast growth factor 23 (FGF23). ATRA, 1,25DHCC, and FGF23 stimulate the formation of cytochrome p enzyme 24 (CYP 24). CYP 24 is the enzyme that is responsible for inactivation of 1,25DHCC and for the formation of 24,25-dihydroxyvitamin D. CYP24 inhibits 1-α hydroxylase, which is needed for the formation of 1,25DHCC

well as stimulation of bone matrix synthesis in mature osteoblasts (Imai et al. 2013, van Beek et al. 2006, Wang et al. 2002).

Effects of vitamin D on hypervitaminosis A related to bone

Most effects on the skeleton attributed to vitamin A are caused by ingestion of crude sources of vitamin A, which also contains other components that might be responsible for bone pathology. This was already discussed by Moore and Wang in 1945. In their study in young rats ingesting high dietary levels of vitamin A (40000 IU / 6 g diet) for 20 days, they also added vitamin D (1 mg / 6 g, for 20 days) to the diet in one group, which developed spontaneous fractures later, compared to the group solely supplemented with vitamin A. Fractures were seen mostly in the center of the diaphysis of long bones; the cortices became thinner and the end of the bones were sometimes ankylosed with the formation of calluses. They also speculated that high vitamin D content of the food might have an antagonistic effect for high vitamin A intake, but that this might be dose dependent, with very high levels of vitamin D not being protective. Cho et al. (1975) and Seawright and Hrdlicka (1974) did not conclude that low Ca, high P intake influenced development of the lesions seen in hypervitaminosis A. Nutritional secondary hyperparathyroidism (NSHP) from raw liver diets was suggested by Watson (1998), which might aggravate bone lesions. Osteoporosis due to NSHP alone often remains subclinical in adult cats. NSHP results in increased bone resorption and replacement of this bone tissue by connective tissue. Parathyroid hyperplasia was noted in leghorn chickens with

93 Part II

high vitamin A intake (660 or 330 IU vitamin A per g of body weight daily for 21 days) with otherwise balanced food (Tang et al. 1985), which can be explained by up-regulation of 24-hydroxylase, which hydroxylates 1,25DHCC and thereby eliminates the negative feedback on PTH synthesis.

A human case report suggested induced bone resorption due to high vitamin A plasma levels in combination with high vitamin D intake. The high vitamin A plasma levels in this case could not be explained by increased dietary intake, but by the hypercalcemia due to hypervitaminosis D, and as a consequence by reduced clearance of vitamin A due to renal failure (Granado-Lorencio et al. 2012). Hypercalcemia also occurs because 6 of the direct, osteoporotic/bone resorptive effects of vitamin A in humans. In dogs with hypercalcemia, a decreased glomerular filtration rate has been noticed, with possibly the same consequences for vitamin A excretion (Fox and Health 1984). In cats, this effect is likely to be absent, as the urinary excretion of vitamin A is already very low (Raila and Schweigert 2002).

Increased osteoblastic activity was noted in rats consuming large amounts of vitamin A and D (Belanger and Clark 1967). In rats vitamin A (probably ATRA, rather than retinol) suppressed PTH levels together with increased 24,25OHD levels (DeLuca 1982), and suppressed 25OHD levels when given in a high dose, but not in a low dose. In vitamin D intoxicated rats, vitamin A lowered Ca and 25OHD plasma levels, demonstrating that high levels of vitamin A interfere with Ca-regulating hormones (Frankel et al. 1985). They also demonstrated that addition of calcitonin, and thus reduced osteoclast activity, reduces the osteoclastic effects of vitamin A on bone.

Conclusions

The effects of hypervitaminosis A maybe the result of an overload of substrate for CYP26A1 and CYP26B1 enzymes, resulting in an intracellular increase in ATRA, especially in osteoblasts and osteoclasts, and subsequent alteration in bone metabolism due to different gene expression. Cats seem to have distinct response to hypervitaminosis A compared to other species. Maybe in cats there is less binding of RANKL, or more osteoprotegrin compared to other species, resulting in less osteoclast activation, and more osteoblast activity, which might explain development of hyperostosis together with osteoporosis. ATRA has proven to decrease binding of proteins to nuclear factor Κβ (Lackey and Hoag 2010), which might explain the reduced RANKL binding. The increased osteoclasia, resulting in osteoporosis demonstrated in several species on a high vitamin A intake, might cause a periostal reaction due to pulling forces on this weakened bone by tendons and ligaments. As vitamin A up-regulates CYP24 (24 hydroxylase) and thus 24,25OHD synthesis that stimulates bone formation without concomitant increase in bone resorption (Boyan et al 2010, van Leeuwen et al 2001), this may lead to the development of skeletal hyperostosis in cats.

94 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

References • Cho, D.Y., Frey, R.A., Guffy, M.M., Leipold, H.W., 1975. Hypervitaminosis A in the dog. • Abu-Abed, S., Dollé, P., Metzger, D., Beckett, American Journal of Veterinary Research 36, B., Chambon, P., Petkovich, M., 2001. The 1597-1603. retinoic acid-metabolizing enzyme, CYP26A1, • Christakos, S., Raval-Pandya, M., Wernyj, is essential for normal hindbrain patterning, R.P., Yang, W., 1996. Genomic mechanisms vertebral identity, and development of involved in the pleiotropic actions of posterior structures. Genes & Development 1,25-dihydroxyvitamin D3. Biochemical 15, 226-240. Journal 316, 361-371. • Al Mheid, I., Patel, R.S., Tangpricha, V., • Clark, L., 1973. Growth rates of epiphyseal Quyyumi, A.A., 2013. Vitamin D and plates in normal kittens and in kittens fed 6 cardiovascular disease: Is the evidence solid? excess vitamin A. A fluorescent labelling and European Heart Journal 34, 3691-3698. histological study. Journal of Comparative • Allegretto, E.A., Shevde, N., Zou. A., Howell, Pathology 83, 447-460. S.R., Boehm, M.F., Hollis, B.W., Pike, J.W., • Clark, L., Seawright, A.A., Gartner, R.J.W., 1995. Retinoid X receptor acts as a hormone 1970. Longbone abnormalities in kittens receptor in vivo to induce a key metabolic following vitamin a administration. Journal of enzyme for 1,25-dihydroxyvitamin D3. The Comparative Pathology 80, 113-120. Journal of Biological Chemistry 270, 23906- • Cline, J.L., Czarnecki-Maulden, G.L., Losonsky, 23909. J.M., Sipe, C.R., Easter, R.A., 1997. Effect of • Bélanger, L.F., Clark, I., 1967. Increasing Dietary Vitamin a on Bone Density Alpharadiographic and histological in Adult Dogs. Journal of Animal Science 75, observations on the skeletal effects of 2980-2985. hypervitaminoses A and D in the rat. • Davies, M.R., Ribeiro, L.R., Downey-Jones, M., Anatomical Record 158, 443-451. Needham, M.R.C., Oakley, C., Wardale, J., 2009. • Benson, A.A., Toh, J.A., Vernon, N., Jariwala, Ligands for retinoic acid receptors S.P., 2012. The role of vitamin D in the are elevated in osteoarthritis and may immunopathogenesis of allergic skin contribute to pathologic processes in the diseases. Allergy 67, 296–301. osteoarthritic joint. Arthritis and Rheumatism • Boyan, B.D., Hurst-Kennedy, J., Denison, T.A., 60, 1722-1732. Schwartz, Z., 2010. 24R,25-Dihydroxyvitamin • DeLuca, H.F., 1982. Metabolism and D3 [24R,25(OH)2D3] controls growth plate mechanism of action of vitamin In: Bone and development by inhibiting apoptosis in the Mineral Research Annual 1 William A. Peck reserve zone and stimulating response to (Ed). Excerpta Medica, Amsterdam. p12-13. 1α,25(OH)2D3 in hypertrophic cells. Journal of • Ettinger, B., Burr, D.B., Ritchie, R.O., 2013. Steroid Biochemistry and Molecular Biology Proposed pathogenesis for atypical femoral 121, 212-216. fractures: Lessons from materials research. • Chattha, K.S., Kandasamy, S., Vlasova, A.N., Bone 55, 495-500. Saif, L.J., 2013. Vitamin A deficiency impairs • Fernández, I., Ortiz-Delgado, J.B., Sarasquete, adaptive B and T cell responses to a prototype C., Gisbert, E., 2012. Vitamin A effects on monovalent attenuated human rotavirus vertebral bone tissue homeostasis in gilthead vaccine and virulent human rotavirus sea bream (Sparus aurata) juveniles. Journal challenge in a gnotobiotic piglet model. PLoS of Applied Ichthyology 28, 419-426. ONE 8, Article number e82966.

95 Part II

• Féron, F., Burne, T.H., Brown, J., Smith, E., • Haruna, A., Kawai, K., Takaba, T., 1992. Dietary McGrath, J.J., Mackay-Sim, A., Eyles, D.W., calcinosis in the cat. Journal of Animal Clinal 2005. Developmental vitamin D3 deficiency Research Foundation 1, 9–16. alters the adult rat brain. Brain Research • Hazewinkel, H.A.W., 1994. Skeletal Disease. Bulletin 65, 141- 148. In: The Waltham Book of Clinical Nutrition of • Fox, J., Heath III, H., 1984. Effect of plasma the Dog & Cat. J.M. Wills & K.W. Simpson Eds., calcium concentration on the metabolic Elsevier, Oxford, England, 413-416. clearance rate of calcitonin in the dog. • Hazewinkel, H.A.W , Tryfonidou, M.A, 2002. Hormone and Metabolic Research 16, 46-49. Vitamin D3 metabolism in dogs. Molecular • Frankel, B.J., Sehlin, J., Täljedal, I.B., 1985. and Cellular Endocrinology 197, 23-33. 6 Vitamin D3 stimulates calcium-45 uptake • Hoogendijk, W.J., Lips, P., Dik, M.G., Deeg, by isolated mouse islets in vitro. Acta D.J., Beekman, A.T., Penninx, B.W., 2008. Physiologica Scandinavica 123, 61-66. Depression is associated with decreased • Freytag, T.L., 2001. Vitamin A metabolism and 25-hydroxyvitamin D and increased toxicity in the domestic cat. Thesis, University parathyroid hormone levels in older adults. California, Davis. Archives of General Psychiatry 65, 508-512. • Freytag, T.L., Liu, S.M., Rogers, Q.R., Morris, J.G., • Hough, S., Avioli, L.V., Muir, H., Gelderblom, D., 2003. Teratogenic effects of chronic ingestion Jenkins, G., Kurasi, H., Slatopolsky, E., Bergfeld, of high levels of vitamin A in cats. Journal M.A., Teitelbaum, S.L., 1988. Effects of of Animal Physiology and Animal Nutrition hypervitaminosis A on the bone and mineral 87, 42-51. metabolism of the rat. Endocrinology 122, • Franch, J., Pastor, J., Franch, B., Durall, I., 2933-2939. Manzanares, M.C., 2000. Back-scattered • How, K.L., Hazewinkel, H.A.W., Mol, J.A., 1994. electron imaging of a non-vertebral case of Dietary vitamin D dependence of cat and hypervitaminosis A in a cat. Journal of Feline dog due to inadequate cutaneous synthesis Medicine and Surgery 2, 49-56. of vitamin D. General and Comparative • Goedegebuure, S.A., Hazewinkel, H.A., 1981. Endocrinology 96, 12-18. Nutrition and bone metabolism. Tijdschrift • Huysseune, A., 2000. Skeletal system. In: The voor Diergeneeskunde 106, 234-242. laboratory fish. G. Ostrander (Ed.). Academic • Granado-Lorencio, F., Rubio, E., Blanco- Press, London, 307-317. Navarro, I., Pérez-Sacristán, B., Rodríguez- • Hyppönen, E., Berry, D.J., Wjst, M., Power, C., Pena, R., García López, F.J., 2012. 2009. Serum 25-hydroxyvitamin D and IgE – a Hypercalcemia, hypervitaminosis A and significant but nonlinear relationship. Allergy 3-epi-25-OH-D3 levels after consumption of 64, 613–620. an “over the counter” vitamin D remedy. a • Imai, Y., Youn, M.-Y., Inoue, K., Takada, I., case report. Food and Chemical Toxicology Kouzmenko, A., Kato, S., 2013. Nuclear 50, 2106-2108. receptors in bone physiology and diseases. • Green, A.S., Tang, G., Lango, J., Klasing, K.C., Physiological Reviews 93, 481-523. Fascetti, A.J., 2012. Domestic cats convert [ • Kneissel, M. , Studer, A., Cortesi, R., Šuša, M., 2H 8]-β-carotene to [ 2H 4]-retinol following a 2005. Retinoid-induced bone thinning is single oral dose. Journal of Animal Physiology caused by subperiosteal osteoclast activity in and Animal Nutrition 96, 689-700. adult rodents. Bone 36, 202-214.

96 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

• Lackey, D.E., Hoag, K.A., 2010. Vitamin A • Moore, T., Wang, Y.L., 1945. Hypervitaminosis upregulates matrix metalloproteinase-9 A. Biochemical Journal 39, 222-228. activity by murine myeloid dendritic cells • Morita, T., Awakura, T., Shimada, A., Umemura, throμgh a nonclassical transcriptional T., Nagai, T., Haruna, A., 1995. Vitamin D mechanism. Journal of Nutrition 140, toxicosis in cats: natural outbreak and 1502-1508. experimental study. Journal of Veterinay • Lascelles, B.D., 2010. Feline degenerative joint Medical Science 57, 831-837. disease. Veterinary Surgery 39, 2-13. • Morris, J.G., 1999. Ineffective vitamin D • Lietz, G., Lange, J., Rimbach, G., 2010. synthesis in cats is reversed by an inhibitor of Molecular and dietary regulation of β,β- 7-dehydrocholestrol-δ7-reductase. Journal of carotene 15,15′-monooxygenase 1 Archives of Nutrition 129, 903-908. 6 Biochemistry and Biophysics 502, 8-16. • Morris, J.G., 2002. Idiosyncratic nutrient • Lind, T., Hu, L., Lind, P.M., Sugars, R., requirements of cats appear to be diet- Andersson, G., Jacobson, A., Melhus, H., 2012. induced evolutionary adaptations. Nutrition Microarray profiling of diaphyseal bone of rats Research Reviews 15, 153-168. suffering from hypervitaminosis A. Calcified • Morris, J.G., Earle, K.E., Andersen, P.A., 1999. tissue international 90, 219-229. Plasma 25-hydroxyvitamin D in growing • Maddock, C.L., Wolbach, S.B., Maddock, S., kittens is related to dietary intake of 1949. Hypervitaminosis A in the dog. The cholecalciferol. Journal of Nutrition 129, Journal of nutrition 39, 117-137. 909-912. • Marill, J. Idres, N., Capron, C.C., Nguyen, E., • Morris, P.J., Salt, C., Raila, J., Brenten, T., Kohn, Chabot, G.G., 2003. Retinoic Acid Metabolism B., Schweigert, F.J., Zentek, J., 2012. Safety and Mechanism of Action: A Review. Current evaluation of vitamin A in growing dogs. Drug Metabolism 4, 1-10. British Journal of Nutrition 108, 1800–1809. • McGrath, J.J., Eyles, D.W., Pedersen, C.B., • Morris-Kay, G., Ward, S., Sokolova, N., 1994. Anderson, C., Ko, P., Burne, T.H., Norgaard- The role of retinoids in normal development Pedersen, B., Hougaard, D.M., Mortensen, P.B., and retinoid-induced malformations. Archives 2010. Neonatal Vitamin D Status and Risk of of Toxicology Supplement 16, 112-117. Schizophrenia: A Population-Based Case- • Nanda, R., van der Linden, F.P.G.M., Jansen, Control Study. Archives of General Psychiatry H.W.B., 1970. Production of cleft palate with 67, 110. dexamethasone and hypervitaminosis A in rat • Mellanby, E., 1921 Experimental rickets. embryos. Experientia 26, 1111-1112. Special report. Series 61. Medical Research • Narbaitz, R., Stumpf, W.E., Sar, M., Huang, Council. London. S., DeLuca, H.F., 1983. Autoradiographic • Mellanby, E., 1938. The experimental localization of target cells for 1α, production of deafness in young animals by 25-dihydroxyvitamin D3 in bones from diet. Journal of Physiology 94, 380-398. fetal rats. Calcified Tissue International 35, • Moore, F.M., Kudisch, M., Richter, K., Faggella, 177-182. A., 1988. Hypercalcemia associated with rodenticide poisoning in three cats. Journal of the American Veterinary Medical Association 193, 1099-1100.

97 Part II

• Norman, A.W., Bishop, J.E., Bula, C.M., • Rodeheffer, C., Von Messling, V., Milot, S., Olivera, C.J., Mizwicki, M.T., Zanello, L.P., Lepine, F., Manges, A.R., Ward, B.J., 2007. Ishida, H., Okamura, W.H., 2002. Molecular Disease manifestations of canine distemper tools for study of genomic and rapid virus infection in ferrets are modulated by signal transduction responses initiated by vitamin A status. Journal of Nutrition 137, 1α,25(OH)2-vitamin D3. Steroids 67, 457-466. 1916-1922. • National Research Council (NRC), 2006. • Ross, A.C., Hammerling, U., 1994. Retinoids Nutrient requirements of dogs and cats. and the immune system. Sporn M.B., Roberts Washington. A.B., Goodman D.S. Eds. New York: Raven • Olson, J.A., 2001. Vitamin A. In: Handbook of Press; The RetinoidsBiology, Chemistry, and 6 Vitamins 3rd ed. Marcel Dekker New York, Medicine. (2nd ed.) pp521–544. NY. Rucker, R.B. Suttie, J.W. McCormick, D.B. • Rothman, K.J., Moore, L.L., Singer, M.R., Machlin, L.J. eds., pp1-50. Nguyen, U.-S.D.T., Mannino, S., Milunsky, A., • Pike, J.W., Meyer, M.B., Bishop, K.A., 2012. 1995. Teratogenicity of high vitamin A intake. Regulation of target gene expression by New England Journal of Medicine 333, 1369- the vitamin D receptor - An update on 1373. mechanisms. Reviews in Endocrine and • Sato, R., Yamagishi, H., Naito, Y., Murakami, Metabolic Disorders 13, 45-55. D., Oshima, K., Takagi, H., Fujita, S., Sasaki, J., • Polizopoulou, Z.S., Kazakos, G., Patsikas, M.N., 1993. Feline vitamin D toxicosis caused by a Roubies, N., 2005. Hypervitaminosis A in the commercially available cat food. Journal of cat: A case report and review of the literature. the Japan Veterinary Medical Association 46, Journal of Feline Medicine and Surgery 7, 577-581. 363-368. • Schwartz, Z., Dean, D.D., Walton, J.K., • Quarles, L.D., 2012. Role of FGF23 in vitamin Brooks, B.P., Boyan, B.D., 1995. Treatment D and phosphate metabolism: Implications of resting zone chondrocytes with in chronic kidney disease. Experimental Cell 24,25-dihydroxyvitamin D3 [24,25-(OH)2D3] Research 318, 1040-1048. induces differentiation into a 1,25-(OH)2D3- • Raila, J., Mathews, U., Schweigert, F.J., 2001. responsive phenotype characteristic of Plasma transport and tissue distribution of growth zone chondrocytes. Endocrinology β-carotene, vitamin A and retinol-binding 136, 402-411. protein in domestic cats. Comparative • Schweigert, F.J., Raila, J., Wichert, B., Kienzle, Biochemistry and Physiology - A Molecular E., 2002. Cats absorb β-carotene, but it is not and Integrative Physiology 130, 849-856. converted to vitamin A. Journal of Nutrition • Raila, J., Schweigert, F.J., 2002. Physiological 132, 1610S-1612S. preculiarities in the vitamin A metabolism • Schweigert, F.J., Raila, J., Harrison, E.H., of carnivores. Tierarztliche Praxis Ausgabe K: Mahmood Hussain, M., 2002. Mechanisms Kleintiere - Heimtiere 30, 1-7. involved in the intestinal digestion and • Reichel, R., Payne, J.H., Otter, A., Holliman, A., absorption of dietary vitamin A. Journal of Foster, A.P., Mitchell, E.S., Jones, J.R., Schock, Nutrition 132, 324-325. A., Woodger, N.G., 2012. Hypervitaminosis A • Seawright, A.A., English, P.B., Gartner, R.J.W., in lambs fed milk replacer. Veterinary Record 1967. Hypervitaminosis A and deforming 171, 448. cervical spondylosis of the cat. Journal of Comparative Pathology 77, 29-39.

98 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

• Seawright, A.A., English, P.B., Gartner, R.J., • Tryfonidou, M.A., Holl, M.S., Stevenhagen, 1970. Hypervitaminosis A of the cat. Advances J.J., Buurman, C.J., Deluca, H.F., Oosterlaken- in veterinary science and comparative Dijksterhuis, M.A., Van Den Brom, W.E., medicine 14, 1-27. Van Leeuwen, J.P.T.M., Hazewinkel, H.A.W., • Seawright, A.A., Hrdlicka, J., 1974. 2003a. Dietary 135-fold cholecalciferol Pathogenetic factors in tooth loss in young supplementation severely disturbs the cats on a high daily oral intake of vitamin A. endochondral ossification in growing dogs. Australian Veterinary Journal 50, 133-141. Domestic Animal Endocrinology 24, 265-285. • Semba, R.D., 2012. The discovery of the • Tryfonidou, M.A., Holl, M.S., Vastenburg, M., vitamins. International Journal for Vitamin Oosterlaken-Dijksterhuis, M.A., Birkenhäger- and Nutrition Research 82, 310-315. Frenkel, D.H., van den Brom, W.E., Hazewinkel, 6 • Senoo, H., Imai, K., Mezaki, Y., Miura, M., Morii, H.A., 2003b. Hormonal regulation of calcium M., Fujiwara, M., Raila, R., 2012. Accumulation homeostasis in two breeds of dogs during of Vitamin A in the Hepatic Stellate Cell of growth at different rates. Journal of Animal Arctic Top Predators. Anatomical Record 295, Science 81, 1568-1580. 1660-1668. • Tryfonidou, M.A., Hazewinkel, H.A.W., Kooistra, • Sih, T.R., Morris, J.G., Hickman, M.A., 2001. H.S., 2010. Calciotropic Hormones. In: Clinical Chronic ingestion of high concentrations of Endocrinology of Dogs and Cats, Rijnberk cholecalciferol in cats. American Journal of A. and Kooistra H.S. Eds., Schlütersche, Veterinary Research 62, 1500-1506. Hannover, Germany, pp256-257. • Spoorendonk, K.M., Peterson-Maduro, J., • Van Beek, E. , Löwik, C., Karperien, M., Renn, J., Trowe, T., Kranenbarg, S., Winkler, Papapoulos, S., 2006. Independent pathways C., Schulte-Merker, S., 2008. Retinoic in the modulation of osteoclastic resorption acid and Cyp26b1 are critical regulators by intermediates of the mevalonate of osteogenesis in the axial skeleton. biosynthetic pathway: The role of the retinoic Development 135, 3765-3774. acid receptor. Bone 38, 167-171. • Tang, K.N., Rowland, G.N., Veltmann Jr., J.R., • Van Leeuwen, J.P., van den Bemd, G.J., van 1985. Vitamin A toxicity: comparative changes Driel, M., Buurman, C.J., Pols, H.A., 2001. in bone of the broiler and leghorn chicks. 24,25-Dihydroxyvitamin D(3) and bone Avian Diseases 29, 416-429. metabolism. Steroids 66, 375-380. • Tryfonidou, M.A., Stevenhagen, J.J., Van Den • Von Gunten, S., Cortinas-Elizondo, F., Kollarik, Bemd, G.J.C.M., Oosterlaken-Dijksterhuis, M.A., M., Beisswenger, C., Lepper, P.M., 2013. Deluca, H.F., Mol, J.A., Van Den Brom, W.E., Van Mechanisms and potential therapeutic targets Leeuwen, J.P.T.M., Hazewinkel, H.A.W., 2002. in allergic inflammation: Recent insights. Moderate cholecalciferol supplementation Allergy: European Journal of Allergy and depresses intestinal calcium absorption in Clinical Immunology 68, 1487-1498. growing dogs. Journal of Nutrition 132, • Wang, X., Wu, J., Shidoji, Y., Muto, Y., Ohishi, 2644-2650. N., Yagi, K., Ikegami, S., Shinki, T., Udagawa, N., Suda, T., Ishimi, Y., 2002. Effects of geranylgeranoic acid in bone: Induction of osteoblast differentiation and inhibition of osteoclast formation. Journal of Bone and Mineral Research 17, 91-100.

99 Part II

• Watson, A.D., 1998. An old, lame cat with reduced food intake and stiff neck. Australian Veterinary Journal 76, 800-801. • Watson, L.C., Marx, C.E., 2002. New onset of neuropsychiatric symptoms in the elderly: possible primary hyperparathyroidism. Psychosomatics 43, 413- 417. • Wiersig, D.O., Swenson, M.J., 1967. Teratogenicity of vitamin A in the canine. Federation Proceedings 26, 486. 6 • Yamamoto, K., Sadahito, K., Yoshikawa, M., Nobuyuki, O., Mikami, O., Yamada, M., Nakamura, K., Yasuyuki, N., 2003. Hyena disease (premature physeal closure) in calves due to overdose of vitamins A, D3, E. Veterinary and Human Toxicology 45, 85-87. • Yanagawa, N., Lee, D.B., 1992. Renal handling of calcium and phosphorus. In: Disorders of bone and mineral metabolism. F.L. Coe and M.J. Favus (eds.) Raven Press, New York pp3-40. • Winaver, J., Sylk, D.B., Teredesai, P.R., Robertson, J.S., Puschett, J.B., 1980. Micropuncture study of the acute renal tubular transport effects of 25-hydroxyvitamin D3 in the dog Mineral and Electrolyte Metabolism 4, 178-188. • Wolf, G., 2004. The Discovery of Vitamin D: The Contribution of Adolf Windaus. Journal of Nutrition 134, 1299-1302. • Wolke, R.E., Nielsen, S.W., Rousseau Jr., J.E., 1968. Bone lesions of hypervitaminosis A in the pig. American Journal of Veterinary Research 29, 1009-1024.

100 The interaction of vitamin A and vitamin D with emphasis on bone metabolism

101 102 Chapter 7

Cutaneous vitamin D synthesis in carnivorous species

Ronald J. Corbee Arie B. Vaandrager Marja J.L. Kik Martijn R. Molenaar Herman A.W. Hazewinkel

Submitted to Journal of Animal Physiology and Animal Nutrition

103 Part II

Abstract

The aim of this study was to investigate the differences of the ability to synthesize sufficient amounts of vitamin D in the skin of different carnivorous species. To this end skin tissue of 22 different carnivorous species were collected from dead animals from zoo’s and our pathology department. Wistar rat skin served as a positive control. Cholesterol, 7-DHC, and vitamin D content was determined after UVB exposure at 37°C, and compared to non-irradiated skin. Overall, there was a significant effect of species and skin thickness, but not of UVB irradiation, on 7-DHC and vitamin D concentrations of the skin. The relatively low cutaneous 7 levels of the vitamin D precursor 7-DHC observed in this study suggest that most terrestrial carnivores are unable to synthesize sufficient amounts of vitamin D. The results have to be taken into account when preparing food for these species when held under captive conditions.

104 Cutaneous vitamin D synthesis in carnivorous species

Introduction

Dogs and cats cannot synthesize sufficient amounts of vitamin D in the skin under the influence of ultraviolet B light (i.e. 280-315nm) (UVB), and thus for their vitamin D requirement depend solely on dietary vitamin D (How et al. 1994). Vitamin D is synthesized by photo isomerization of 7-dehydrocholesterol (7-DHC) into pre-vitamin D, followed by heat isomerization into vitamin D (Figure 1). This physicochemical process has been revealed both in vivo as in vitro (How et al. 1994, Morris 1999). The amount of 7-DHC in rat skin proved to be much higher compared to dog and cat skin (How et al. 1994), explaining the lower ability of dogs and cats to synthesize sufficient amounts of vitamin D in the skin. Vitamin D status in man and animals is evaluated by determining plasma 25-hydroxy vitamin D (25OHD) levels. Puppies raised on synthetic vitamin D deficient food, meeting dietary calcium and phosphorus requirements, developed radiological 7 and histological diagnosed rickets and low plasma levels of 25OHD, despite daily UVB irradiation of the skin (Hazewinkel and Tryfonidou 2002). Furthermore, puppies raised on identical food only differing in vitamin D content (11.4 vs. 1350μg of vitamin D3 per kg diet), the high vitamin D group had 75 times higher 25OHD plasma levels compared to the low vitamin D group. In kittens, plasma 25OHD levels are also dependent on dietary intake and linearly correlate with the amount of vitamin D in the diet, regardless of exposure to sunlight (Morris et al. 1999). A reverse in the cutaneous vitamin D synthesis capacity was demonstrated in cats that were supplemented with an inhibitor of delta 7-DHC reductase, suggesting that over-expression of this enzyme shifts the vitamin D precursor 7-DHC to a different metabolic pathway (i.e. cholesterol synthesis) (Morris 1999). This led to the conclusion that carnivorous species could lose the capacity of cutaneous vitamin D synthesis, since they can cover their requirements with the vitamin D content present in their prey. To the authors’ knowledge, nothing is known about the ability of cutaneous vitamin D synthesis for other carnivorous species, which is important for possibly required adaptations of daily rations of these carnivorous species in captivity. Therefore, the aim of this study was to investigate the differences of the ability to synthesize sufficient amounts of vitamin D in the skin of different carnivorous species.

Materials and Methods

Skin tissue (3x3cm) of different carnivorous species were collected from the back of dead animals from zoo’s and our pathology department, and stored at -70°C for further analysis. Laboratory rat skin (8 months old male Wistar) served as a positive control. Cholesterol, 7-DHC, and vitamin D content was determined after UVB exposure at 37°C, and compared to non-irradiated skin. Two pieces of 1x1cm skin were cut from the sample, and weighed (Table 1). The pieces were thawed and heated to 37°C in a water bath for 2 hours prior to extraction (UVB- group), or UVB exposure (UVB+ group) and extraction. All skin samples were freed of subcutaneous tissue, shaved, and the UVB+ samples were irradiated with UVB (Arcadia 12% UVB D3 reptile lamp 15W, peak at 305 nm, Arcadia, Redhill, UK) light for 30 min (only the UBV+ group). Irradiation of the skin samples of the

105 Part II

UVB+ group was done at 1200±100μW/cm2 (=12.00±1.00 J/s.m2). With a total exposure time of 1800 seconds, consequently irradiation corresponds with 21600 J/m2 = 2.160 J per cm2. The amount of UVB administered was measured by the use of an UVB-meter (Solarmeter 6.2UV meter, Solartech Inc., USA). After 2 hours at 37°C, the pieces were cut into very small chunks and transferred into glass tubes and the lipids were extracted by the method as described by Bligh and Dyer (1959). In short 0.8 mL millipure water and 3 mL chloroform:methanol 1/2, v/v, were added to the tubes and were mixed for 40 min by regular vortexing. Then the tubes were centrifuged for 2 min at 2000 rpm. The supernatant was transferred to new glass tubes and 2 mL millipure water and 2 mL chloroform were added. After vortexing for 30 sec, the tubes were centrifuged for 5 min at 2000 rpm. The lower layer was transferred to a new tube and evaporated under N2 gas. These samples were stored at -20°C prior to MS-analysis. Before MS-analysis, the samples 7 were resuspended in 500 µL chloroform:methanol 1/1, v/v containing 0.002% butylated hydroxytoluene (BHT) as an anti-oxidant, and 20 µL was injected on a Lichrospher RP18-e column. A gradient was generated from acetonitrile:water 95/5, v/v, to acetone/ chloroform 85/15, v/v, at a constant flow rate of 1 mL/min. Total run time per sample was 13 min. MS of lipids was performed using Atmospheric Pressure Chemical Ionization (APCI) on a Biosystems API-4000 Q-trap (MDS Sciex, Concord, ON, Canada). The system was controlled by Analyst version 1.4.2 software (MDS Sciex, Concord, ON, Canada) and operated in positive ion mode and in the multiple reaction monitoring (MRM) mode using the following settings: source temperature 420°C, nebulizer gas (GS1) 5, nebulizer current 3 μA, curtain gas 10, collision gas. High and declustering potential and collision energy were empirically optimized for each compound. The MRM transitions (m/z) used were: 367.3 → 159.1 for 7-DHC, desmosterol and vitamin D (species were identified with regard to retention times 7.2, 7.5, and 6.45 min, respectively), and 369.3 → 287.3 for cholesterol.

Data analysis was performed using Analyst 1.4.2 software (MDS Sciex, Concord, ON, Canada). Quantitation was done relative to standards run separately (all steroid standards were from Sigma-Aldrich (St. Louis, MO, USA)). As cholesterol is expected to be extracted with similar efficiency as vitamin D and 7-DHC, and cholesterol is a good indicator of the amount of cellular material present in the skins, the data are expressed as a ratio to cholesterol. All skin samples were analysed in duplicate. In case of high variations (i.e. >20%) the skin samples were analysed for another two times. The average levels of two samples are demonstrated.

Statistical analysis Multivariate analysis was performed to determine effects of UVB, species, and skin thickness (determined by the weight of the piece of skin of 1x1cm) on skin 7-DHC, and skin vitamin D concentrations, the latter two expressed per nmol cholesterol.

106 Cutaneous vitamin D synthesis in carnivorous species

Skin Lumisterol or tachysterol Fig. 1: Metabolic pathways of 7-dehydrocholesterol (7-DHC). UVB Pre-vitamin D can be synthesized from 7-DHC in the skin under the Cholesterol 7-DHC Pre-vitamin D influence of ultraviolet B light Desmosterol (UVB). Under the influence of UVB, Skin temperature pre-vitamin D can be converted vitamin D into lumisterol or tachysterol. Under influence of the skin Blood temperature, pre-vitamin D can be DBP-vitamin D DBP converted into vitamin D, which is bound to vitamin D binding protein (DBP) in the blood. If 7-DHC is not 7 used for vitamin D synthesis, it can be converted to cholesterol and desmosterol

Table 1: 7-dehydrocholesterol (7-DHC) and vitamin D (VitD) concentration of the skin samples before (-) and after (+) exposure to UVB (2.16 J per cm2), expressed in pmol per nmol cholesterol. The weight of the skin (Sw) is given in mg per cm2

Species Latin name 7-DHC - 7-DHC + VitD - VitD + Sw Rat Rattus norvegicus 270 256 3 4 177 Gray wolf Canis lupus 11 6 10 11 464 Dog Canis lupus familiaris 7 7 11 7 117 African wild dog Lycaon pictus 32 30 4 5 186 Fox Vulpes vulpes 30 27 6 3 245 Arctic fox Vulpus lagopus 35 20 1 1 481 Polar bear Ursus maritimus 27 27 1 1 239 Red panda Ailurus fulgens 0 11 7 7 165 European badger Meles meles 0 0 1 4 425 European polecat Mustela putorius 12 25 6 2 487 Otter Lutra lutra 24 23 6 7 350 Raccoon Procyon 32 28 1 1 130 Seal Phoca vitulina 164 123 1 1 1186 Ringed seal Pusa hispida 108 106 1 1 389 Californian sea lion Zalophus califonianus 176 378 8 13 1492 Lion Panthera leo 6 6 5 7 124 Ocelot Leopardus pardalis 16 17 2 3 186 Bobcat Lynx rufus 21 21 1 1 88 Fishing cat Prionailurus viverrinus 5 6 9 10 75 Cat Felis catus 16 15 1 1 55 Yellow mongoose Cynictis penicillata 12 7 1 10 150 Meerkat Suricata suricatta 254 188 1 2 136

107 Part II

Results

A total of 44 skin samples from 22 different species (Figure 2) were analyzed and 7-DHC and vitamin D concentrations of the skin in pmol expressed per nmol of cholesterol before and after irradiation with UVB are demonstrated in Table 1. Wistar rat skin, as non- carnivorous control, contained the highest amount of 7-DHC. Most carnivorous species had relatively low, but detectable levels of 7-DHC in their skin, although the meerkat and sea-carnivores had higher levels of 7-DHC, comparable to those of rats (Table 1). The relatively low levels of 7-DHC seemed not to be related to a lower cholesterol synthesis as, the levels of desmosterol, another precursor of cholesterol were not dissimilar between rat and the various carnivorous species (result not shown). All carnivorous species had detectable levels of vitamin D in their skin before and after irradiation with UVB. After 7 UVB irradiation, rat, African wild dog, gray wolf, meerkat, otter, and yellow mangoose skin 7-DHC levels decreased, together with an increasing level of vitamin D. In the European polecat skin vitamin D levels dropped after UVB exposure, which coincided with an increase in 7-DHC. In fishing cat, ocelot, red panda, and sea lion skin 7-DHC levels also increased after UVB exposure. The European badger has higher levels of vitamin D in the skin after UVB exposure without a detectable decrease of 7-DHC. The variation in the effects of UVB irradiation might be real, but, due to the limited sample number, we cannot exclude the possibility that it reflects biological variation between the samples. Overall, there was a significant effect of species and skin thickness, but not of UVB irradiation, on 7-DHC and vitamin D concentrations of the skin.

Discussion

Up till now, 7-DHC concentration of the skin is regarded as the indicator for sufficient cutaneous vitamin D synthesis (Kohler et al. 2013). The 21 carnivorous species investigated here differ in their 7-DHC content of the skin, and all of them had lower 7-DHC content of the skin compared to the omnivorous rat. Most carnivorous species are thus unlikely to be able to synthesize sufficient amounts of vitamin D in their skin. From the investigated carnivorous species, meerkat and sea-carnivores had the highest levels of 7-DHC in their skin. The presence of this vitamin D precursor in higher levels in the skin compared to other carnivores could enable these species to form sufficient amounts of vitamin D in the skin. Synthesis of vitamin D from 7-DHC in the skin demands photo isomerization by UVB irradiation to pre-vitamin D, followed by heat isomerisation into vitamin D (Kasian et al. 2012). This final step requires several hours at body temperature and does not require the presence of UVB. When pre-vitamin D is not heat-isomerized it can be reconverted in 7-DHC, and subsequently converted to lumisterol or tachysterol (Kasian et al. 2012). The theoretical potential of large amounts of 7-DHC to form vitamin D after long sun exposure is not realized due to further isomerization of pre-vitamin D by these reactions as well as by degradation of vitamin D by UVA (Jablonski and Chaplin 2013). In vivo, the vitamin D formed in the skin is readily bound to abundantly present vitamin D binding proteins for transportation to the target organs (Tian et al. 1994, Hazewinkel and

108 Cutaneous vitamin D synthesis in carnivorous species

Canis Canis lupus

Canis lupus familiaris

Canidae Lycaon Lycaon pictus

Vulpes vulpes 7

Vulpes Vulpes lagopus

Caniformia

Ursoidea Ursidea Ursus maritimus Meles Meles meles

Ailuridae Ailures fulgens Mustela Mustela putorius

Arctoidea Musteloidea Mustelidea Mustelinae Lutra Lutra lutra

Lutrinae

Procyonidae Procyon Phoca vitulina

Pinnipedia Phocidae Phocinae Phocini Pusa hispida Carnivora Ottariidea Ottoriinae Zalophus Zalopus californianus

Feloidae Feledae Pantherinae Panthera leo

Leopardus Leopardus pardalis

Lynx Lynx rufus

Feliformia Prionailurus Prionailurus viverrinus

Felis Felis catus

Viverroidea Herpestoidea Herpestoidea Cynictis Cynictis penicillata

Suricata Suricata suricatta

Fig. 2: Phylogenetic tree of the species used in this study

109 Part II

Tryfonidou 2002). When the vitamin D is not taken up by this mechanism, as was not the case in this in vitro study, it might be that the synthesized vitamin D was reconverted into 7-DHC, or further photo isomerized to lumisterol or tachysterol (Kasian et al. 2012). This can explain the decrease of vitamin D and increase of 7-DHC as was demonstrated in skin samples of some of the investigated carnivorous species. The dosage of UBV used in our study (2.16 J per cm2) is similar to the UVB dosage of How et al. (1994) (2.25 J per cm2), excluding differences in findings due to different dosage of UVB. No effective vitamin D synthesis was demonstrated in the skin in most carnivorous species after UVB exposure, while in rat skin significant vitamin D synthesis was demonstrated. This does not rule out the possibility of cutaneous vitamin D synthesis by carnivores, because it might be that the skin of the carnivores is thicker or more pigmented (Libon et al. 2013), and therefore needs longer exposure to UVB or a higher amount of UVB to become effective. 7 However, exposure of the skin samples with the highest weight and pigmentation (i.e. Californian sea lion and seal) for 2 hours did not result in an increase in skin vitamin D content (data not shown). In humans, the main function of epidermal pigmentation is protection of DNA against UVB (Jablonski and Chaplin 2013). In the natural habitat, the amount of UVB and the amount of vitamin D in the diet may be different from current housing conditions in zoo animals, which might result in vitamin D deficiency (Pye et al. 2013). Body temperature also influences isomerization, which may be different from the standardized 37°C that we used in our study. It is not known what the diet was of the animals that were used in this study. Diet may also affect 7-DHC concentration of the skin before UVB exposure due to the need for cholesterol, as well as after UVB exposure by the reconversion of the vitamin D present before UVB irradiation in the skin into 7-DHC, as was possibly the case in European polecat, dog, and fox skin in this study. Cholesterol content of the skin influences 7-DHC concentration, as was demonstrated by up-regulation of 7-DHC reductase in human fibroblast cultures in a cholesterol deficient medium (Wassif et al. 1998). However we did not find differences in desmosterol levels, the other cholesterol precursor. From human studies it is known that ageing lowers the 7-DHC concentration in the skin (Gallagher et al. 2013). Elderly people and people are therefore more prone to vitamin D deficiency and need longer sun exposure for adequate synthesis of vitamin D. We did not know the age of the animals used in this study, so we can only speculate on the influence of ageing. From the skin samples only parts of the skin from the back were available. In chickens, 7-DHC concentrations were 30 times higher in leg skin compared to back skin. After UVB exposure of the whole body (0.5 J per cm2), pre-vitamin D was only present in uncovered skin of legs and feet, while no pre-vitamin D was demonstrated in back skin (Tian et al. 1994). In dairy cows and sheep, just like in humans, the whole skin is capable of vitamin D synthesis, even when it is covered with hair, fur, or wool (Hymøller and Jensen 2010). Feathers of chickens provide protection to UVB, while the pigmented skin in the legs and feet are capable of vitamin D synthesis. Although we assumed that in carnivores there are no differences as a result of their fur, we shaved all the skin samples prior to measurement. Because of all the variables that influence cutaneous vitamin D synthesis (both in vivo and in vitro) as were presented in this study, it is hard to draw firm conclusions from this in vitro study. It seems unlikely

110 Cutaneous vitamin D synthesis in carnivorous species

that terrestrial carnivores are capable of synthesizing sufficient amounts of vitamin D in their skin due to the low levels of 7-DHC. Furthermore, the 7-DHC concentration in the skin of carnivores found in this study are 4-40 times lower compared to skin of sheep and goats (Kohler et al. 2013). In vivo studies, using vitamin D deficient diets are needed to prove if the carnivorous species are able to synthesize vitamin D in the skin of biological significance as demonstrated in dogs (Hazewinkel and Tryfonidou 2002) and cats (Morris 1999).

Conclusion

This in vitro study demonstrated that carnivorous species differ in 7-DHC and vitamin D concentrations in the skin. The relatively low cutaneous levels of the vitamin D precursor 7-DHC observed in this study suggest that most terrestrial carnivores are unable to 7 synthesize sufficient amounts of vitamin D. The results have to be taken into account when preparing food for these species when held under captive conditions.

111 Part II

References • Kohler, M., Leiber, F., Willems, H., Merbold, L., Liesegang, A., 2013. Influence of altitude on • Bligh, E.G., Dyer, W.J., 1959. A rapid method vitamin D and bone metabolism of lactating of total lipid extraction and purification. sheep and goats. Journal of Animal Science Canadian Journal of Biochemistry and 91, 5259-5268. Physiology 37, 911–917. • Libon, F., Cavalier, E., Nikkels, A.F., 2013. Skin • Gallagher, J.C., Peacock, M., Yalamanchili, color is relevant to vitamin D synthesis. V., Smith, L.M., 2013. Effects of vitamin D Dermatology 227, 250-254. supplementation in older African American • Morris, J.G., 1999. Ineffective vitamin D women. Journal of Clinical Endocrinology and synthesis in cats is reversed by an inhibitor of Metabolism 98, 1137-1146. 7-dehydrocholestrol-δ7-reductase. Journal of • Hamilton, J.G., Comai, K., 1988. Rapid Nutrition 129, 903-908. 7 separation of neutral lipids, free fatty acids • Morris, J.G., Earle, K.E., Andersen, P.A., 1999. and polar lipids using pre-packed silica Sep- Plasma 25-hydroxyvitamin D in growing Pak columns. Lipids 23, 1146-1149. kittens is related to dietary intake of • Hazewinkel, H.A.W, Tryfonidou, M.A., 2002. cholecalciferol. Journal of Nutrition 129, Vitamin D3 metabolism in dogs. Molecular 909-912. and Cellular Endocrinology 197, 23-33. • Pye, G.W., Ellis, W., Fitzgibbon, S., Opitz, B., • How, K.L., Hazewinkel, H.A.W., Mol, J.A., 1994. Keener, L., Hollis, B.W., 2013. Serum vitamin Dietary vitamin D dependence of cat and D levels in free-ranging koalas (Phascolarctos dog due to inadequate cutaneous synthesis Cinereus). Journal of Zoo and Wildlife of vitamin D. General and Comparative Medicine 44, 480-483. Endocrinology 96, 12-18. • Retra, K., Bleijerveld, O.B., Gestel, R.A. van, • Hymøller, L., Jensen, S.K., 2010. Vitamin D3 Tielens, A.G.M., Hellemond, J.J. van, Brouwers, synthesis in the entire skin surface of dairy F.F., 2008. A simple and universal method cows despite hair coverage. Journal of Dairy for the separation and identification of Science 93, 2025-2029. phospholipid molecular species. Rapid • Jablonski, N.G., Chaplin, G., 2013. Epidermal Communications in Mass Spectrometry 12, pigmentation in the human lineage is an 1853-1862. adaptation to ultraviolet radiation. Journal of • Tian, X.Q., Chen, T.C., Lu, Z., Shao, Q., Human Evolution 65, 671-675. Holick, M.F., 1994. Characterization of the • Karsten, K.B., Ferguson, G.W., Chen, T., Holick, translocation process of vitamin D3 from the M.F., 2009. Panther chameleons, furcifer skin into the circulation. Endocrinology 135, pardalis, behaviorally regulate optimal 655-661. exposure to UV depending on dietary vitamin • Wassif, C.A., Maslen, C., Kachilele-Linjewile S., D status. Physiological and Biochemical Lin, D., Linck, L.M., Connor, W.E., Steiner, R.D., Zoology 82, 218-225. Porter, F.D., 1998. Mutations in the human • Kasian, N.A., Vashchenko, O.V., Gluhova, Y.E., sterol delta 7-reductase gene at 11q12-13 Lisetski, L.N., 2012. Effect of the vitamin D cause Smith-Lemli-Opitz syndrome. American photosynthesis products on thermodynamic Journal of Human Genetics 63, 55-62. parameters of model lipid membranes. Biopolymers and Cell 28, 114-120.

112 Cutaneous vitamin D synthesis in carnivorous species

113 114 Chapter 8

Composition and use of puppy milk replacers in German Shepherd puppies in the Netherlands

Ronald J. Corbee Marianna A. Tryfonidou Irene P. Beckers Herman A.W. Hazewinkel

Journal of Animal Physiology and Animal Nutrition 2012; Volume 96, Issue 3, Pages 395-402

115 Part II

Abstract

Enostosis or eosinophilic panosteitis is a common disease in young growing large-breed dogs, such as the German Shepherd, and the risk of developing the disease by 3-4 months of age is increased by a high calcium intake. The aim of the study was to investigate whether German Shepherd puppies raised on milk replacers receive more calcium and/or vitamin D than their requirements in the pre-weaning period and thus are at increased risk of developing skeletal diseases. To this end, we surveyed German Shepherd breeders in the Netherlands about the use of puppy milk replacers (PMR). The metabolizable energy (ME), calcium, phosphorus and vitamin D content of the eight most-used PMR was compared with that of bitch milk, as reported in the literature. 8 The protein and fat content of most PMR were somewhat lower (range 24.4g to 33.2g per 100g on dmb and 18.3g to 37.5g per 100g on dmb, respectively) compared to bitch milk (31.9g and 40.2g on dmb, respectively). The vitamin D content of one of the PMR samples was 7-fold the level recommended by the NRC (2006) and 3-fold the average level of bitch milk. The clinical relevance of this high amount is questionable, as bitch milk contains mainly 25-hydroxy-vitamin D (3843μg (96.1 IU) per 100g on dmb) and only limited amounts of vitamin D (524μg (13.3 IU) per 100g on dmb), as was determined in this study. Dutch German Shepherd breeders tended to overfeed their puppies. We calculated that misguided use of PMR can increase the risk of excessive calcium, phosphorus and possibly vitamin D intake during a vulnerable period, potentially giving rise to bone and cartilage problems later in life.

116 Composition and use of puppy milk replacers in German Shepherd puppies in the Netherlands

Introduction

High calcium intake can result in enostosis and disorders of endochondral ossification, manifested as osteochondrosis (Schoenmakers et al. 2000a, Schoenmakers et al. 2000b, Slater et al. 1992). A dietary intake of calcium of 1.5% on a dry matter basis (dmb) showed there were no skeletal deformities in growing dogs (Weber et al. 2000). The current recommended upper limit during the period of early growth (<14 weeks of age) is 1.6% calcium on a dmb (FEDIAF 2008, Laflamme 2000, NRC 2006). A 3-fold increase of calcium intake in partially weaned puppies increased the risk of enostosis at 3-4 months of age (Schoenmakers et al. 2000a). It has been demonstrated that growing dogs cannot down-regulate calcium absorption and that a high dietary calcium intake results in a proportionally increased calcium uptake (Tryfonidou et al. 2002b). High calcium intake in turn elicits a high postprandial release of calcitonin (Hazewinkel et al. 1991). Large-breed puppies fed on a diet with 3.3% calcium (dmb) between 3-6 weeks of age developed hypercalcitoninism and C-cell hyperplasia at latter age when compared with control 8 puppies from the same litters raised on a diet with 1.1% calcium (dmb) (Schoenmakers et al. 1999). Calcitonin inhibits osteoclast activity and thus bone remodelling. As a result the canaliculi in the cortex cannot adapt to the growth in blood vessel diameter, resulting in congestion both inside the medullary cavity and subperiosteally. The latter results in enostosis with shifting lameness caused by the painful lifting of the periosteum from the underlying bone.

Tryfonidou et al. (2002a) showed that a dietary vitamin D content of 25μg (1000IU) per 1000 kcal Metabolizable Energy (ME) (which is 8 times the level recommended by the NRC (2006)) can lead to disturbances of endochondral ossification, with normal calcium, phosphorus and 1,25-dihydroxy-cholecalciferol blood concentrations, when fed to large- breed puppies of 6-21 weeks of age. According to recent NRC and European Pet Food Industry Federation (FEDIAF) recommendations, the safe upper limit for dietary vitamin D levels in the early growth period is 20μg (or 800IU) per 1000 kcal ME (FEDIAF 2008, NRC 2006). A vitamin D intake of 1000-fold the safe upper limit due to consumption of over-supplemented PMR resulted in vitamin D intoxication with poor development and condition, impaired mobility and severe calcifications in the kidneys in Airedale Terrier puppies (Kamphues et al. 1990).

The aim of this study was to investigate whether German Shepherd puppies, when raised on PMR, receive levels of calcium or vitamin D higher than their requirements in the pre-weaning period and are thus at risk for developing skeletal diseases. To this end, we surveyed Dutch German Shepherd breeders about their use of PMR and compared the ME, calcium, phosphorus, and vitamin D content of eight most often used commercially available PMR with that of bitch milk, as reported in the literature. Since there is no data available about the normal vitamin D content in bitch milk, we determined vitamin D and 25-hydroxy-vitamin D concentrations in bitch milk during mid-lactation.

117 Part II

Materials and methods Survey among breeders A cohort (Appendix 1) of 184 non-professional breeders in the Netherlands was received from the German Shepherd breeders club and consisted of 23 questions about their breeding experience and the use of PMR. The amount of PMR fed by the breeders was calculated as grams of milk powder, based on the amount fed and the mixing ratio recorded by the breeder or, if not available, the mixing ratio recommended by the manufacturer. Puppy body weight was recorded daily up to 3 weeks of age, and weekly thereafter. A preliminary questionnaire was tested prior to being used and necessary modifications were made. The questions were closed with predetermined and pre-coded categories.

Analysis of commercially available PMR and bitch milk PMR used most frequently by the breeders in our survey were obtained directly from 8 the manufacturers to prevent possible differences in shelf time. Two samples were taken from the same lot from each product for further analysis. Bitch milk was obtained from lactating bitches of 4 different breeds (1 Labrador Retriever, 1 Stabyhoun, 1 Beagle and 1 Beagle-Bedlington Terrier cross) on day 10 of lactation. Per dog 2 samples of 20 ml were taken and stored at -20ºC until further analysis. The Weende analysis and spectrophotometric determination of the calcium (NEN_EN_ ISO6869) and phosphorus (ISO6491) content of PMR was performed by the Premervo Research Laboratory, Utrecht (the Netherlands). The moisture content was determined by drying at 80ºC under pressure of up to 100 Torr for 4 hours according to 71/393/EEG. Crude protein was determined by Kjeldahl method and calculated according to ISO 5983-2. Crude fat was determined after acid hydrolysis according to 98/64/EG, method B. Crude fiber was determined by NEN_EN_ISO6865. Crude ash was determined by 71/250/ EEG. The vitamin D and 25-hydroxy-vitamn D content of the PMR and bitch milk were measured by high-performance liquid chromatography (HPLC) analysis, performed by TNO (The Netherlands Organisation for Applied Scientific Research) in Zeist, as described elsewhere (Antalick et al. 1977, Berg et al. 1986, Liu Bo et al. 2003). The intra- and inter assay coefficient of variation were 5.7% and 6.6%, respectively for vitamin D and 15.2% and 6.1%, respectively for 25-hydroxy-vitamin D. The levels were compared with the known reference values of bitch milk as reported in the literature (Anderson et al. 1991, Lonnerdal et al. 1981, Meyer and Zentek 2005, Oftedal 1984).

Determination of calcium and vitamin D intake during the weaning period The daily energy requirements were calculated in kcal ME using the equation 211.9 kcal ME per kg body weight per day (Meyer and Zentek 2005) based on the mean body weight of all the puppies on day 1 and day 14. The energy derived from the PMR was calculated using the unmodified Atwater coefficients (NRC 2006) (e.g. kcal ME = 4x Crude protein + 9x Crude fat + 4x Carbohydrates). The calcium intake from PMR when fed at energy requirement level was calculated by dividing the energy requirement of the puppy (kcal ME) by the energy provided by PMR (kcal ME per 100g on dmb) and multiplied with

118 Composition and use of puppy milk replacers in German Shepherd puppies in the Netherlands

the calcium content of the PMR (grams per 100g on dmb). The vitamin D intake by the puppies from PMR was calculated likewise, i.e. by dividing the energy requirement of the puppy by the energy provided by PMR and multiplied by the vitamin D content of the PMR (IU per 100g on dmb). The calcium and vitamin D intakes of PMR when fed to the puppies at the actual amount were calculated by dividing the maximum amount PMR fed in grams dry matter per day by 100 and multiplying this with the calcium or vitamin D content per 100g on dmb, respectively.

Amount puppy milk replacer fed (grams per day)

100

80 y

60 8

ed in grams per da 40 Amount f 20

0 0 246810 12 14 16 18

Age in days

Fig. 1: Amount of puppy milk replacer in grams per day fed to German Shepherd puppies by the breeders at different ages

9000

8000

7000

6000

5000 Maximum weight eight (g) 4000 Average weight

Body w Minimum weight 3000

2000

1000

0 0 14 28 42 56

Age in days Fig. 2: Body weight of German Shepherd puppies at different ages

119 Part II

Results Breeding experience and the use of PMR by breeders Of 184 breeders, 114 reported using PMR (62%). A complete survey was obtained from 57 breeders, 47% of whom had more than 15 years of experience as breeder; the other 53% had 1-15 years of experience with a median of 12 years. On average the breeders raised 1.8 litters a year. Eight brands of PMR were mainly used. Most breeders (85%) purchased PMR from pet stores, and 5% from veterinary practices. The remainder of PMR was obtained from other sources (for example from the manufacturer). The reasons to feed PMR were delivery by caesarean section, prolonged whelping, agalactia, oligolactia, illness of the bitch, rejection of puppies by the bitch, and impaired growth of the puppies. In addition, most breeders used PMR to gradually wean the puppies from bitch milk to puppy food (i.e. 6 weeks after birth). In total, 76% of the breeders used PMR only during the first 3 weeks: 83% of the breeders used PMR during the first week and 56% during the second week; 20% continued using PMR after the puppies were 3 weeks old. Of the 8 breeders, 4% used PMR only in puppies after the third week of age. Bottle-feeding was the main feeding route. Two breeders occasionally used tube feeding if necessary. The amount of PMR fed varied, depending on the purpose of PMR use, the body weight and the age of the puppies. Remarkably, the breeders could not provide us any protocols about their use of PMR at different ages or at different body weights. Of the breeders, 50% determined the amount fed at their own discretion, 21% used the guidelines of the manufacturer, of which only 8% adapted the amount fed on the basis the puppy’s body weight and growth rate, and 29% fed as much as the puppy was able to drink. Figure 1 shows the PMR amount fed by the breeders, regardless of the purpose for the use of PMR. The amount fed, given in this figure, is the amount that was eaten by the puppies.

Body weight and energy requirements of the puppies The weight curves of 442 puppies (230 males, 212 females) from 41 breeders (74 litters) are given in Figure 2. These data were obtained from these breeders and includes only the puppies that were weighed on the 5 consecutive time points. Data from puppies that died within the 56-days-period were excluded. The body weight of the puppies at different ages and their corresponding energy requirements are given in Table 1.

Analysis of PMR The results of Weende analysis and the calcium, phosphorus and vitamin D content of the PMR are presented (expressed per 100 g dry matter PMR) in Table 2. The ME of PMR varied between 465 and 567.6 kcal per 100g on dmb compared to 553.2 kcal per 100g on dmb in bitch milk. The protein and fat content of most PMR were somewhat lower (range 24.4g to 33.2g and 18.3g to 37.5g per 100g on dmb) compared to bitch milk (31.9g and 40.2g per 100g on dmb). The calcium content of PMR varied between 0.88g and 1.15g per 100g on dmb which is similar to the calcium content of bitch milk (1.01g per 100g on dmb). The phosphorus content of PMR was a bit lower (between 0.58g and 0.83g per 100g on dmb) compared to bitch milk (0.93g per 100g on dmb). The calcium:phosphorus ratio of PMR varied between 1.1 and 1.5 compared to 1.1 – 1.8, as reported in bitch milk (Meyer and

120 Composition and use of puppy milk replacers in German Shepherd puppies in the Netherlands

Table 1: Body weight of the puppies at different ages and the energy requirements calculated according to Meyer and Zentek 2005 (211.9 kcal ME per kg BW per day)

Age in days Average body weight Energy requirements in grams in kcal ME per day 0 520±72 110.2±15.3 14 1217±191 257.9±40.5 28 2860±570 660.0±120.8 56 6280±1120 1330.7±237.3

Table 2: Composition of puppy milk replacers (per 100g dry matter)

Sample C Protein C Fat C Fibre C Ash Energy (kcal Ca P Vitamin D ME) 1 33.2 37.5 0.37 4.6 567.6 0.97 0.67 220 8 2 29.0 22.8 0.33 6.4 487.5 0.92 0.81 300 3 25.0 23.5 0.53 6.2 490.6 0.88 0.58 140 4 30.6 23.5 0.28 5.5 494.4 0.89 0.75 230 5 29.1 23.8 0.14 6.3 493.2 0.91 0.78 350 6 27.9 31.2 0.26 5.6 532.6 0.90 0.83 240 7 24.4 18.3 0.22 6.4 465.0 1.15 0.77 270 8 26.8 26.8 0.54 6.4 506.2 1.11 0.83 190 Bitch milk 31.9 40.2 n/a 5.1 553.2 1.01 0.93 n/a

Crude protein, Crude fat, Crude fibre, Crude ash, calcium, and phosphorus in grams Vitamin D in International Units Energy in kcal ME calculated by Atwater coefficients by NRC 2006 Values from bitch milk derived from Oftedal 1984 (moisture content 77.3%) n/a = not available

Table 3: Calcium intake from the analyzed PMR and from bitch milk when fed at energy requirements and at the maximum amount according to the breeders at day 1 and 14 days

At energy requirements At maximum amount 1 day 14 days 1 day 14 days Bitch milk 0.209 0.450 Sample 1 0.189 0.441 0.533 0.800 Sample 2 0.208 0.486 0.499 0.749 Sample 3 0.197 0.462 0.478 0.717 Sample 4 0.198 0.464 0.483 0.725 Sample 5 0.204 0.476 0.949 0.741 Sample 6 0.186 0.436 0.492 0.738 Sample 7 0.273 0.638 0.620 0.931 Sample 8 0.242 0.566 0.6605 0.907 Upper limit 0.441 1.551

121 Part II

Zentek 2005). The vitamin D content of PMR varied between 140IU and 350IU per 100g on dmb. No 25-hydroxy-vitamin D could be demonstrated in the PMR.

Calcium and vitamin D intake when raised on PMR The calcium intake from PMR and bitch milk (Meyer and Zentek 2005) are given in Table 3. Comparison of the absolute amount of calcium of PMR fed as described by the breeders on days 1 and 14 with that of bitch milk as sole source of nutrition shows that already by day 1 the calcium intake was 3-fold higher with PMR (Table 3). The vitamin D intake from PMR and bitch milk (this study using the sum of 25-hydroxy-vitamin D and vitamin D) are given in Table 4. When PMR was supplied according to the pup’s energy requirements, the vitamin D content of only one brand of PMR exceeded the safe upper limit of 20 μg (or 800 IU) per 1000 kcal ME (FEDIAF 2008, NRC 2006). However, when PMR was given in the maximum amounts as reported by the breeders, only one brand (sample 3) could be regarded as safe on day 1, and 3 brands (samples 1, 3 and 8) on day 14, according to NRC 8 2006 recommendations (Table 4). Analyses of the vitamin D content of the eight PMR showed that measured levels were different from the levels stated by the manufacturer on the PMR label (Table 5).

Vitamin D content of bitch milk The concentrations of vitamin D in bitch milk were on average 1200 μg (30 IU) (range 1050-1300 μg) per L and of 25-hydroxy-vitamin D on average 8800 μg (220 IU) (range 8000-9000 μg) per L. Although bitch milk of German Shepherds was not available, breed differences were not observed. To compare these results with the vitamin D content of PMR, we have to calculate this to μg (IU) per 100g on dmb. The dry matter content of the bitch milk was 22.6% on average (range 21.8% to 23.2%), thus the concentrations of vitamin D and 25-hydroxy-vitamin D in bitch milk were 524μg (13.3 IU) and 3843μg (96.1 IU) per 100g on dmb, respectively, or 974μg (24 IU) and 6947μg (174 IU) per 1000 kcal ME, respectively.

Discussion

The aim of this study was to investigate whether German Shepherd puppies raised on milk replacer receive more calcium or vitamin D than their daily requirements in the pre- weaning period and are thus potentially at increased risk of developing skeletal diseases. Our survey showed that most Dutch German Shepherd breeders use PMR for different purposes, which as a consequence influences the amount of PMR given. When PMR was fed as addition to bitch milk, we were unable to draw conclusions on total intake, as the amount of bitch milk consumed was not measured. The amount fed to orphan puppies was mostly determined by the breeder’s experience instead of according to the manufacturer’s guidelines. There was a tendency to overfeed the puppies, especially because some breeders allowed the puppies to take PMR ad libitum instead of in the amount recommended on the basis of the puppy’s nutritional requirements.

122 Composition and use of puppy milk replacers in German Shepherd puppies in the Netherlands

Table 4: Vitamin D intake from the analyzed PMR and from bitch milk when fed at energy requirements and at the maximum amount according to the breeders at the age of 1 day and 14 days

At energy requirements At maximum amount 1 day 14 days 1 day 14 days Bitch milk 22 51 Sample 1 43 100 121 181 Sample 2 68 159 163 244 Sample 3 32 73 76 114 Sample 4 52 120 125 187 Sample 5 78 183 190 285 Sample 6 49 116 131 197 Sample 7 64 149 146 219 Sample 8 42 97 104 155 8 Upper limit 88 206

Vitamin D in IU per day Upper limit according to FEDIAF 2008

Table 5: Vitamin D declaration of the manufacturer compared with Vitamin D determination by HPLC

Manufacturer HPLC Maximum allowance 320 320 Sample 1 150 220 Sample 2 150 300 Sample 3 200 140 Sample 4 200 230 Sample 5 200 350 Sample 6 150 240 Sample 7 150 270 Sample 8 160 190

Vitamin D in IU per 100g dry matter Maximum allowance according to NRC 2006

123 Part II

The weight curves revealed a huge variance among the puppies, with the heaviest puppies being almost twice as heavy as the lightest puppies. However, this difference was already seen at an early age and can be attributed to genetic or pre-natal differences and does not necessarily indicate an overfed state. The weight curves were similar to those of puppies of other breeds as have been reported in studies by others (Meyer and Zentek 2005).

Puppies that were raised on PMR only, according to the energy requirements for growing puppies as determined by Meyer and Zentek (2005), had a normal calcium intake relative to bitch milk and their calcium intake can be regarded as safe. However, some breeders overfed their puppies, so that the calcium intake exceeded the recommended upper limit (FEDIAF 2008, NRC 2006), mainly in the first 3 weeks of life. For example, feeding PMR sample 7 ad libitum could lead to a calcium intake that is 3-fold higher on a daily basis than that obtained by getting only bitch milk. The dietary phosphorus intake would be increased proportionally. Earlier research has shown that a 3-fold increased calcium 8 intake during the period of partial weaning (i.e., 3-6 weeks of age) increases the risk of developing enostosis at 3-4 months of age (Schoenmakers et al. 1999, Tryfonidou et al. 2002b). In addition, a 3-fold increased calcium intake during the post weaning period increases the risk of developing disturbances in endochondral ossification and thus of skeletal development (Hazewinkel et al. 1985, Hazewinkel et al. 1991, Schoenmakers et al. 2000a). Furthermore, during the period of rapid growth in large-breed dogs, an increased intake of both calcium and phosphorus increases the risk and severity of disturbances in endochondral ossification, in comparison with an increased intake of calcium alone (Schoenmakers et al. 2000a).

With respect to the vitamin D content of PMR, we found a range of 140-350 IU per 100g on dmb. In order to compare the vitamin D content of PMR with normal concentrations in bitch milk we searched through the literature without results. We therefore compared the vitamin D content of PMR with the recommended upper limit and adequate intake from FEDIAF and NRC. The vitamin D content was high in 4 of the 8 PMR samples, with sample 5 exceeding the maximum levels advised by the NRC and exceeding adequate intake by 7-fold (FEDIAF 2008, NRC 2006). All PMR samples except sample 3 exceeded the recommended upper limit if the puppies were being overfed, with a calculated daily intake of 5.2-7.9 μg per day (at 14 days of age) which is almost 15 times the adequate intake. Tryfonidou et al. (2002a) demonstrated in large breed puppies that a daily intake of 5.2 μg vitamin D per day (for a puppy of 1217g BW; i.e. 4.27 μg vitamin D per kg BW per day) caused disturbances in endochondral ossification when this is given from 6-21 weeks of age. The vitamin D content declared on the label is the minimum level that should be analyzable in the product at the end of shelf life, and thus does not accurately reflect the actual content at analysis. Furthermore, 6 (75%) of the labels of PMR studied here only state the amount of vitamin D added to the constituents of the PMR, which explains differences found between the results of vitamin D analysis and the claimed vitamin D content as declared on the PMR label.

124 Composition and use of puppy milk replacers in German Shepherd puppies in the Netherlands

To our knowledge, this is the first report on the vitamin D content in bitch milk. There are no studies on the possible effects of the dietary vitamin D intake of the bitch on the vitamin D content of bitch milk. However, it is well known in man that the maternal dietary vitamin D content has a strong effect on vitamin D status of the infant (Thandrayen and Pettifor 2010, Viljakainen et al. 2010). Interestingly we found that bitch milk contains mainly 25-hydroxy- vitamin D and almost no non-hydroxylated vitamin D, contrarily to the composition of human milk with a vitamin D : 25-hydroxy-vitamin D ratio of 1:1 (Ala-Houhala et al. 1988). Human infants reveal to have 25-hydroxylase activity (Salle et al. 2000). We speculate that the relatively high amounts of 25-hydroxy-vitamin D content of bitch milk may serve for the relative inability of newborn puppy livers to convert vitamin D into 25-hydroxy- vitamin D due to a (relative) deficiency in 25-hydroxylase activity. Further studies are warranted to underscore such a speculation. The influence of excessive dietary vitamin D intake by the bitch on the vitamin D content of bitch milk and consequently on the skeletal development of the puppies when given during the first 3 weeks of life, have not yet been studied. These studies are crucial in order to determine the dietary requirements 8 of vitamin D for lactating bitches and growing puppies before the period of partial weaning. The total vitamin D content of bitch milk (i.e. vitamin D and 25-hydroxy-vitamin D) was twice the recommended intake by NRC (2006). This indicates that the total vitamin D intake from bitch milk is higher than assumed. Since the influence of dietary vitamin D intake by the bitch might influence the vitamin D content of the milk it is important to take into account that all four bitches in our study were fed a diet with an analysed average of 1000 IU vitamin D3 per kg on dmb or 270 IU per 1000 kcal ME, respectively. The bitches were all fed the same diet, from the same batch, by the principal author according to energy requirements. As long as it has not been determined if puppies have adequate hepatic 25-hydroxylase activity, the relevance of excessive vitamin D intake with PMR and the unexpectedly high 25-hydroxy-vitamin D content of bitch milk is unknown. The adequacy of all PMR, lacking 25-hydroxy-vitamin D, is till then questionable.

Excessive calcium intake during the second 3 weeks of life results in panosteitis at older age (i.e. 3-4 months), as revealed by a well-controlled study in Great Danes (Schoenmakers et al. 2000a). We found that 17% of breeders were overfeeding their puppies, and these puppies may thus be at risk for excessive calcium intake from the PMR. A prospective study under well-controlled conditions, is needed to determine whether a high intake of calcium, phosphorus, and possibly vitamin D via PMR in the first 3 weeks of life indeed increases the risk of enostosis and osteochondrosis in young large-breed dogs later in their life. When feeding orphan puppies with PMR only, PMR should be given according to the puppy’s energy requirements to prevent overnutrition and excessive mineral intake, which may adversely affect skeletal development.

Acknowledgements

The authors thank Lennart Oosterbaan and Cor Schreuder from Premervo, The Netherlands for analyzing the ME, calcium, and phosphorus content of the PMR.

125 Part II

References • Laflamme, D.P., 2000. Effect of breed size on calcium requirements for puppies. • Ala-Houhala, M., Koskinen, T., Parviainen, Compendium on Continuing Education for M.T., Visakorpi, J.K., 1988. 25-Hydroxyvitamin the Practicing Veterinarian 23, 66-69. D and vitamin D in human milk: Effects of • Liu Bo, Lang ZhaoBin, Yin XiaoHong, Zhou supplementation and season. American XuDong, Wang, K., 2003. Determination of Journal of Clinical Nutrition 48, 1057-1060. vitamins A and D in dairy products using high • Anderson, R.S., Carlos, G.M., Robinson, I.P., performance liquid chromatography. China Booles, D., Burger, I.H., Whyte, A.L., 1991. Zinc, Dairy Industry 31, 36-37. copper, iron and calcium concentrations in • Lonnerdal, B., Keen, C.L., Hurley, L.S., Fisher, bitch milk. Journal of Nutrition 121, S81-S82. G.L., 1981. Developmental changes in the • Antalick, J.P., Debruyne, H., Faugere, J.G., 1977. composition of beagle dog milk. American Measurement of vitamin D in foods by high Journal of Veterinary Research 42, 662-666. pressure liquid chromatography. Annales de • Meyer, H., Zentek, J., 2005. Ernährung des 8 Falsifications et de l’Expertise Chimique Food Hundes. Parey Verlag, Stuttgart, Germany. Science and Technology Abstracts. • NRC, 2006. Nutrient Requirements of Dogs • Berg, H. van den, Boshuis, P.G., Schreurs, and Cats. In: National Academy Press, W.H.P., 1986. Determination of Vitamin D in Washington DC. dairy products by High Performance Liquid • Oftedal, O.T., 1984. Lactation in the dog: milk Chromatography. Journal of Agricultural and composition and intake by puppies. Journal Food Chemistry 34, 264-268. of Nutrition 114, 803-812. • FEDIAF, 2008. Nutritional Guidelines for • Salle, B.L., Delvin, E.E., Lapillonne, A., Bishop, Complete and Complementary Pet Food for N.J., Glorieux, F.H., 2000. Perinatal metabolism Cats and Dogs. European pet food industry of vitamin D. American Journal of Clinical federation, Bruxelles. Nutrition 71, 1317S-1324S. • Hazewinkel, H.A.W., Brom, W. E. van den, • Schoenmakers, I., Nap, R.C., Mol, J.A., Klooster, A.Th.van ‘t, Voorhout, G., Wees, Hazewinkel, H.A.W., 1999. Calcium A. van, 1991. Calcium metabolism in Great metabolism: an overview of its hormonal Dane dogs fed diets with various calcium and regulation and interrelation with skeletal phosphorus levels. Journal of Nutrition 121, integrity. Veterinary Quarterly 21, 147-153. S99-S106. • Schoenmakers, I., Hazewinkel, H.A.W., • Hazewinkel, H.A.W., Goedegebuure S.A., Voorhout, G., Carlson, C.S., Richardson, D., Poulos P.W., Wolvekamp W.Th.C., 1985. 2000a. Effect of diets with different calcium Influences of different calcium intakes and phosphorus contents on the skeletal on calcium metabolism and skeletal development and blood chemistry of growing development in young Great Danes. Journal great danes. Veterinary Record 147, 652-660. of the American Animal Hospital Association • Schoenmakers, I., Mol, J.A., Hazewinkel, 21, 337-391. H.A.W., 2000b. Hormonal calcium regulation • Kamphues, J., Meyer, H., Pohlenz, J., Wirth, and calcium setpoint in offspring of bitches W., 1990. Vitamin D intoxication in Airedale with different calcium intakes during puppies fed with milk replacer. Kleintierpraxis pregnancy. Journal of Animal Physiology and 35, 458. Animal Nutrition 83, 1-14.

126 Composition and use of puppy milk replacers in German Shepherd puppies in the Netherlands

• Slater, M.R., Scarlett, J.M., Donoghue, S., Kaderly, R.E., Bonnett, B.N., Cockshutt, J., Erb, H.N., 1992. Diet and exercise as potential risk factors for osteochondritis dissecans in dogs. American Journal of Veterinary Research 53, 2119-2124. • Thandrayen, K., Pettifor, J.M., 2010. Maternal vitamin D status: Implications for the development of infantile nutritional rickets. Endocrinology and metabolism clinics of North America 39, 303-320. • Tryfonidou, M.A., Stevenhagen, J.J., Van Den Bemd, G.J.C.M., Oosterlaken-Dijksterhuis, M.A., Deluca, H.F., Mol, J.A., Van Den Brom, W.E., Van Leeuwen, J.P.T.M., Hazewinkel, H.A.W., 2002a. 8 Moderate cholecalciferol supplementation depresses intestinal calcium absorption in growing dogs. Journal of Nutrition 132, 2644-2650. • Tryfonidou, M.A., Van Den Broek, J., Van Den Brom, W.E., Hazewinkel, H.A.W., 2002b. Intestinal calcium absorption in growing dogs is influenced by calcium intake and age but not by growth rate. Journal of Nutrition 132, 3363-3368. • Viljakainen, H.T., Saarnio, E., Hytinantti, T., Miettinen, M., Surcel, H., Mäkitie, O., Andersson, S., Laitinen, K., Lamberg-Allardt, C., 2010. Maternal vitamin D status determines bone variables in the newborn. The Journal of clinical endocrinology and metabolism 95, 1749-1757. • Weber, M., Martin, L., Dumon, H., Biourge, V., Nguyen P., 2000. Growth and skeletal development in two large breeds fed 2 calcium levels. Journal of Veterinary Internal Medicine 14, 388.

127 128 Chapter 9

Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

Ronald J. Corbee Marianna A. Tryfonidou Guy C.M. Grinwis Claudia F. Wolschrijn Seng Fong Lau Ben M.C. Gorissen Arie B. Vaandrager Herman A.W. Hazewinkel

Submitted to The Veterinary Journal

129 Part II

Abstract

The aim of this study was to investigate whether vitamin D supplementation of the standard puppy diet given to puppies after weaning can prevent the development of medial coronoid disease (MCD) by stimulating endochondral ossification, including terminal differentiation of chondrocytes, and mineralization of the cartilaginous template of the developing medial cornoid process (MCP). A litter of Labrador puppies was on purpose bred by mating a dam and a sire with MCD; these dogs are known to produce offspring with MCD. The puppies received a diet supplemented with 50,000 IU vitamin D per kg as fed. Development of MCD was monitored by computed tomography (CT) and plain radiographs every two weeks, and post-mortem by microCT, necropsy, histology, and immunohistochemistry. The results of the vitamin D supplemented puppies were compared to the normally fed 9 control group from a previous litter from the same parent animals and identical investigated. Vitamin D supplementation did result in increased plasma levels of 25-vitamin D and 1,25-vitamin D, but did not prevent development of MCD in growing Labradors. Instead, vitamin D supplementation resulted in increased collagen X staining of the MCP and irregular costal growth plates, demonstrating disturbed endochondral ossification.

130 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

Introduction

Medial coronoid disease (MCD) is a common heritable disease in young large-breed dogs and is characterized by fissures and/or fragmentation of the medial coronoid process (MCP), cartilage, and/or subchondral bone (Fitzpatrick et al. 2009). It is regarded as one of the developmental abnormalities responsible for elbow dysplasia and has a high incidence in Labrador retrievers (Van Ryssen and van Bree 1997, Lavrijsen et al. 2012). Ad libitum feeding, overweight, heavy exercise, and heavy playing are considered risk factors for MCD (Sallander et al. 2006), and while the exact cause of MCD is still unknown, it is likely to be multifactorial (Danielson et al. 2006). As overweight and exercise are risk factors for MCD, uneven stress on the elbow may be an important factor in MCD development (Davidson et al. 2008). Such stress may arise as a result of differences in the length of the radius and ulna, due to differences in longitudinal bone growth, or physiological incongruity during loading (Davidson et al. 2008, House et al. 2009). MCD development has also been related to abnormalities of subchondral bone: histological investigations have revealed micro-cracks in the axial border of the MCP to be more severe at the fragmented site of the MCP than in the rest of the MCP (Danielson et 9 al. 2006). MCD development has also been attributed to disturbed endochondral ossification (Olsson and Reiland 1978). A recent study showed an enlarged calcifying zone (which is the deepest cartilage layer, and is characterized by hypertrofic chondrocytes surrounded by calcifying intercellular substance forming the template of cancellous bone formation) and retained hyaline cartilage in subchondral cancellous bone in the MCP of MCD-positive, but not MCD-negative, Labrador retriever siblings of the same age (Lau et al. 2013b). The enlarged calcifying zone seen in MCD-positive dogs is indicative of disturbed endochondral ossification due to delayed or abnormal terminal differentiation of chondrocytes, and/or delayed calcification of the matrix (Danielson et al. 2006). Similar lesions, i.e. a larger layer of persistent hypertrophic chondrocytes surrounded by uncalcified matrix due to delayed mineralization are seen in osteochondrosis but also in case of rickets (Thompson 2007). Althogether this indicates that susceptibility to MCD development may be related to relative vitamin D deficiency.

Vitamin D influences skeletal development and endochondral ossification by stimulating the terminal differentiation of chondrocytes and the mineralization of cartilage and the newly formed osteoid (Hazewinkel and Tryfonidou 2002). Dogs are unable to synthesize sufficient amounts of vitamin D in the skin under the influence of UVB light and are entirely dependent on dietary vitamin D (How et al. 1994). At birth, dogs have limited body reserves of vitamin D in the liver but are thereafter dependent on their dam’s milk or milk replacer as a source of vitamin D (Corbee et al. 2012). The aim of this study was to investigate whether vitamin D supplementation of the standard puppy diet given to puppies after weaning can prevent the development of MCD. We hypothesized that supplementation with vitamin D has a protective effect by stimulating endochondral ossification, including terminal differentiation of chondrocytes, and mineralization of the cartilaginous template of the developing MCP.

131 Part II

Materials and methods

This study was approved by the Utrecht University Institutional Committee on the Care and Use of Experimental Animals and was in compliance with national legislation on laboratory animal use (DEC 2012.III.07.068).

A litter of Labrador puppies was on purpose bred by mating a dam and a sire with MCD; these dogs are known to produce offspring with MCD. The weight of the puppies was recorded daily during the first 3 weeks of life, and every other day thereafter. The puppies were partially weaned at 3 weeks of age and completely weaned when the puppies were 7 weeks old, using a normal weaning diet (Royal Canin Large dog starter; Table 1) supplemented with 50,000 IU of cholecalciferol per kg of food on an ‘as-fed’ basis for the entire period of the study.

Blood samples were drawn when the puppies were 5 weeks old and then every 2 weeks until euthanasia at 18 weeks of age, or earlier if MCD was diagnosed by computed tomography 9 (CT). Blood samples were collected by jugular venipuncture into ethylenediaminetetra- acetic acid (EDTA)-coated tubes stored on melting ice. The tubes were centrifuged at 4° C at 3000 rpm and plasma was stored at -20° C until further analysis. Plasma concentrations of total calcium (Ca), phosphate (P), total protein, and albumin were determined with a Unicel DXC-600 (Beckman Coulter Inc., Brea, CA, USA). The plasma concentration of 25-hydroxyvitamin D (25-vitamin D) was determined by high-performance liquid chromatography-mass spectrometry (HPLC-MS) after hexane extraction. To this end, 0.25 mL demineralized water, 0.50 mL ethanol, 150 pmol D6-25-vitamin D as internal standard, and 4 mL hexane containing 0.002% butylated hydroxytoluene as antioxidant were added to 0.2 mL plasma. Samples were vortexed vigorously 3 times for 5 seconds and then centrifuged (5 min, 1000g). The upper phase was transferred to a new glass tube and hexane was evaporated under nitrogen gas. The lipid fraction was stored at -20° C until it was reconstituted in 0.1 mL methanol/chloroform (1/1, v/v). HPLC separation was performed as described by Testerink et al. (2012). Briefly, 10 µL of the lipid fraction was injected onto a Lichrospher RP18-e column (5 μm, 250 x 4.6 mm; Merck, Darmstadt, Germany) and eluted with an acetonitrile:water 95/5 to acetone/chloroform 85/15, v/v gradient at a constant flow rate of 1 mL/min. Total run time per sample was 13 min. Mass spectrometry of lipids was performed using Atmospheric Pressure Chemical Ionization (APCI) on a Biosystems API-4000 Q-trap (MDS Sciex, Concord, ON, Canada). The system was controlled by Analyst version 1.4.2 software (MDS Sciex, Concord, ON, Canada) and operated in positive ion mode and in the multiple reaction monitoring (MRM) mode using the following settings: source temperature 420° C, nebulizer gas (GS1) 5, nebulizer current 3 μA, curtain gas 10, collision gas high and declustering potential and collision energy were empirically optimized for each compound. The MRM transition (m/z) used was: 383.3 → 365.3 for 25-vitamin D, and 389.3 → 371.3 for D6-25-vitamin D. 25-vitamin D was measured relative to the internal standard. The exact concentration of the internal standard was determined before each experiment by spectrophotometry (D6-25-vitamin

132 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

Table 1: Chemical analysis of the food given to puppies in the control group and in the vitamin D-supplemented group

Per 100 g as fed Per 100 g dry matter Per 100 kcal ME Crude protein (g) 30 32.6 6.8 Crude fat (g) 22 23.9 5 NFE (g) 31.3 34 7.1 Crude fiber (g) 6.4 7 1.4 Moisture (g) 8 Calcium (g) 1.2 1.3 0.27 Phosphorus (g) 0.95 1.03 0.21 Sodium (g) 0.4 0.43 0.09 Chloride (g) 0.58 0.63 0.1 Potassium (g) 0.7 0.76 0.16 Magnesium (g) 0.07 0.08 0.016 EPA+DHA (g) 0.3 0.33 0.07 9 Vitamin A (IU) 2100 2283 474 Vitamin D (IU) 120 130 27.1 Vitamin E (mg) 60 65 13.5

Metabolizable Energy (ME) kcal / 100 g 443 482 kJ / 100 g 1854 2016

The vitamin D supplemented group received a supplement containing 5000 IU of vitamin D per mL Per 100 g of food, 1 mL of the supplement was added NFE = nitrogen free extract, EPA = eicosapentaenoic acid, DHA = docosahexaenoic acid

D at 264 nm using a molar absorption coefficient (M-1.cm-1) of 19400. The recovery was approximately 80% for the internal standard, and the intra-assay variation was 8.5% for 25-vitamin D. Plasma concentrations of 1,25-dihydroxyvitamin D (1,25-vitamin D) were determined after immuno-extraction in a competitive radioimmunoassay (IDS AA-54F1, EliTech Systems Pvt Ltd., Gujarad, India) according to the manufacturer’s instructions. Samples were diluted 6x with assay buffer. The limit of detection was 20 pmol/L, and the inter-assay variation was 10% at 95 pmol/L.

CT and plain radiographs of the elbow joints were taken under general anesthesia every 2 weeks from the age of 5 weeks (n=3) or 6 weeks (n=2) onward to monitor the development of the bony structures of the secondary ossification centers (Appendix A, Voorhout and Hazewinkel 1987) and of the MCP of the elbow joints. CT was performed

133 Part II

as described previously (Lau et al. 2013a) with a third-generation single-slice helical CT scanner (Philips Secura, Philips, Eindhoven, the Netherlands) using 120 kV and 120 mA with an exposure time of 1000 ms. Radiographs of the elbow joints were made with a digital radiography system (Philips digital Rad TH, Philips, Eindhoven, the Netherlands) using 50 kV and 8 mA. Puppies that developed MCD, as revealed on CT images of one or both elbow joints, were euthanized with an IV overdose of barbiturate immediately after diagnosis. The remaining puppies were euthanized at 18 weeks of age, which was defined as the study endpoint, as most MCD-positive puppies in similar studies developed MCD before that age (Lau et al. 2013a, b). Immediately after euthanasia, the elbow joints and the growth plates of both 9th ribs were collected, examined macroscopically, and fixed in 4% neutral buffered formaldehyde.

Thereafter, micro-CT was performed to visualize the MCP and to identify the exact location of fissures or fragments of the MCP, when present, as described previously (Lau et al. 2013). In brief, the ulnas were scanned in a microCT system (Scanco 80, Scanco Medical, Switzerland) from 2 cm proximal to the MCP to 2 cm distal to the MCP, using 55 9 kV, 145 mA, an exposure time of 1 s, and a pixel size of 36.9 µm.

After micro-CT, the MCP and costal growth plates were decalcified in 10% EDTA for 6 weeks, and then dissected and embedded in paraffin. Histological evaluation was performed on hematoxylin and eosin (HE)-, and safranin O-stained tissue sections. Immunohistochemistry was performed for collagen type X as described previously (Lau et al. 2013b). Collagen type X is a short chain collagen produced specifically by hypertrophic chondrocytes during endochondral ossification. Histomorphometry was performed with the aid of Image J software (v1.47b Colour Convolution, NIH). The surface area of the four zones of the developing joint cartilage, i.e. tangential, transitional, radial, and calcifying zones, and the relative surface area positive for type X collagen was determined as described previously (Lau et al. 2013b). The mean width and standard deviation of each costal growth plate were measured at 12 different pre-defined locations. The percentage standard deviation of the mean growth plate width (%sd) indicated the irregularity of the width of the growth plate (Tryfonidou et al. 2002). The ratio between the width of the hypertrophic zone and the proliferation zone of the costal growth plate was calculated to demonstrate differences in the stage of endochondral ossification.

The effect of vitamin D supplementation on skeletal development was established by comparing data from vitamin D-supplemented puppies with earlier reported data (Lau et al. 2013b) obtained in MCD-positive and MCD-negative puppies from two different litters of puppies with the same parents and raised under the same housing conditions and on the same standardized diet (containing 27.1 IU vitamin D per 100 kcal metabolizable energy) as the research animals described in this study.

Statistical analyses were performed with the use of R-statistics (R i386 3.0.1). All data were tested for normal distribution by a Kolmogorov-Smirnov test. Normally distributed

134 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors data are presented as means±sd. Data that were not normally distributed are presented as medians and range. Multivariate analysis was performed to correct for possible interactions between time points, sex, and diet group. A Fisher test was performed to correct for multiple comparisons; a p-value of <0.05 was set as the level of significance.

Results

The control group consisted of 14 puppies, 9 males and 5 females, from two litters. Seven puppies (5 males and 2 females) developed MCD. The youngest MCD-positive dog in the control group was 15 weeks old at the time of diagnosis based on CT. A third litter of 2 males and 3 females were raised on the vitamin D supplemented food. In these puppies, MCD was first detected by CT in a male puppy at 16 weeks of age and in a female puppy at 18 weeks of age, similar to the ages of puppies with MCD in the control group (Table 2). The other 3 puppies (1 male, 2 females) did not have CT-demonstrable MCD by 18 weeks of age. The puppies in the control group and in the vitamin D-supplemented group had similar growth curves and none were lame during the study period. Vitamin D intake was 57 IU per kg0.75 in the control group and 2427 IU per kg0.75 in the vitamin D-supplemented 9 group (Table 1).

Plasma Ca, P, total protein, and albumin levels in the vitamin D-supplemented group did not differ from those of the control group with mean values during the study period for total Ca 2.85±0.06 mmol/L, P 3.15±0.18 mmol/L, total protein 60±4 g/L, and for albumin 31±2 g/L. While plasma levels of 25-vitamin D and 1,25-vitamin D were comparable in the two groups at the start of the study, 25-vitamin D and 1,25-vitamin D plasma levels were approximately twice as high in the supplemented group as in the control group (Figure 1). The development of secondary ossification centers was similar up to 18 weeks of age in the two groups and did not differ between MCD-positive and MCD-negative puppies (Appendix A).

At necropsy, the supplemented puppy that was MCD-positive at 16 weeks of age had an incomplete fissure in the articular cartilage of the right, but not left, elbow. The supplemented puppy that was MCD-positive at 18 weeks of age had an incomplete fissure in the articular cartilage of the left, but not right, elbow. Similar findings were seen in the MCD-positive puppies in the control group. The elbow joints of all MCD-negative puppies were normal in both groups (Table 2).

Micro-CT revealed bilateral fragmentation of the MCP of the 16-week-old supplemented puppy and bilateral fissure of the MCP in the 18-week-old supplemented puppy (Figure 2). The remaining puppies in the vitamin D-supplemented group did not have any abnormalities at the study endpoint (i.e. 18 weeks of age). Representative examples of micro-CT imaging, histology, and immunohistochemistry findings for the MCD-positive and MCD-negative puppies of the supplemented and control groups are shown in Figure 2. The calcifying zone of the joint cartilage at the side of the MCP in the MCD-positive

135 Part II

Table 2: Necropsy findings for puppies in the control group and in the vitamin D-supplemented group Dog Age Sex Weight Limb CT Necropsy MicroCT Histology (Weeks) (kg) (Cartilage) (Mineralization) Vitamin D supplemented group 1 16 M 13.5 L Suspect No fissure Non-mineralized zone MCD R Positive Inc. Fissure Non-mineralized area MCD 2 18 F 22.8 L Positive Inc. Fissure Non-mineralized zone MCD R Suspect No fissure Non-mineralized zone MCD 3 18 F 23.0 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 4 18 M 19.0 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 5 18 F 21.0 L Negative No fissure Normal Normal R Negative No fissure Normal Normal Control group 1 15 F 14.8 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 2 15 F 16.2 L Suspect No fissure Non-mineralized zone MCD 9 R Negative No fissure Normal Normal 3 16 M 16.6 L Suspect No fissure Non-mineralized zone MCD R Suspect No fissure Non-mineralized area MCD 4 17 F 20.5 L Suspect No fissure Non-mineralized zone MCD R Suspect No fissure Non-mineralized zone MCD 5 18 M 19.2 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 6 18 M 19.1 L Positive Fissure Non-mineralized area MCD R Positive Inc. fissure Non-mineralized area MCD 7 19 M 21.5 L Positive No fissure Non-mineralized zone MCD R Positive Inc. Fissure Non-mineralized area MCD 8 19 M 22.8 L Negative Fissure Non-mineralized area MCD R Negative Inc. Fissure Non-mineralized area MCD 9 25 M 28.6 L Suspect Fissure Non-mineralized area MCD R Suspect Fissure Non-mineralized area MCD 10 25 M 28.4 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 11 26 M 28.4 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 12 26 M 29.4 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 13 27 F 24.4 L Negative No fissure Normal Normal R Negative No fissure Normal Normal 14 27 M 33.6 L Negative No fissure Normal Normal R Negative No fissure Normal Normal

Sex is male (M) or female (F). Limb is left (L) or right (R). CT findings are negative, suspicion of medial coronoid disease (Suspect), or positive for medial coronoid disease. Macroscopic evaluation at necropsy of the cartilage did not detect fissures, incomplete fissures (Inc. Fissure), or complete fissures (Fissure). Micro-computed tomography (MicroCT) revealed a normal medial coronoid process (Normal), a non-mineralized line within the mineralized area (Non-mineralized zone), or a non-mineralized area between two mineralized areas (Non-mineralized area)

136 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

Fig. 1: 25-vitamin D and 1,25-vitamin D plasma concentrations of puppies in the vitamin D-supplemented group and the control group (* p<0.05, ** p<0.01)

Fig. 2: Comparison of the proximal view of the 3-D reconstructed micro-computed tomography (MicroCT) images of the medial coronoid process with corresponding (cut at yellow line), hematoxylin and eosin, Safranin O, and collagen X staining in medial coronoid disease negative (MCD-) and medial coronoid disease positive (MCD+) joints. Representative examples of tissue from control puppies and vitamin D-supplemented puppies of the same age (18 weeks old)

137 Part II

Fig. 3: Lesion within the cartilage template of the developing medial coronoid process in a 16-week-old puppy with medial coronoid disease. Safranin O (A and C), collagen X staining (B), and Hematoxylin and eosin (D). Note that the section was taken from the beginning of the fissure. The designated rectangles are magnified and flipped in horizontal plane. The lesion is located in bone and cartilage, and some pycnotic nuclei in the degenerative cartilage cells, vacuolization, and debris of both cartilage and subchondral bone (arrow) can be seen. Degeneration of the chondrocytes surrounding the defect can also be seen (diminished Safranin O staining)

Fig. 4: Representative examples of regular (dog from the control group) and irregular (dog from the vitamin D-supplemented group) growth plates from the 9th ribs (collagen X staining)

138 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

Table 3: Histology findings for control (Control) and vitamin D-supplemented (vitamin D+) puppies with (+) or without (–) medial coronoid disease (MCD)

Control Vitamin D+ MCD- MCD+ MCD- MCD+ Surface area cartilage MCP (mm2) 2.0±0.2A 2.6±0.4B 2.0±0.2A 2.5±0.4B p=0.019 Width of cartilage layer MCP 0.79 2.07 1.00 (mm) (0.75-0.81)A (1.67-2.77)B (0.82-1.02)A (2.21-2.63)B p=0.013 Collagen X ratio MCP (%) 0.02 1.43 0.83 (0.00-2.55)A (0.00-11.83)B (0.21-3.60)C (9.90-11.42)D p<0.05 Growth plate (9th rib) width 0.83 0.81 (mm) (0.78-1.02)A N/A (0.69-1.02) (0.81-1.17) NS Growth plate irregularity (%sd) 10±1A N/A 10±1A 17±9B p=0.041 Ratio hypertrophy zone / 0.77 0.74 proliferation zone (9th rib) (0.63-0.84) N/A (0.65-0.87) (0.72-0.81) NS 9 A significantly different from B,C,D (p<0.05)

Fig. 5: The relative surface area positive for Collagen type X staining of the medial coronoid process (MCP) in medial coronoid disease (MCD) positive (+) and negative (-) puppies in the control group (Control) and the vitamin D-supplemented group (VitaminD+). Each dot represents one MCP

139 Part II

puppies was significantly thicker than that of the MCP in the MCD-negative puppies in both groups (Figure 2, Table 3). In the two supplemented puppies with MCD, the fissure was located at the junction of the cartilage and the subchondral bone (Figure 2), with dead chondrocytes being seen in the calcifying zone of the cartilage (Figure 3), as was also seen in the MCD-positive puppies of the control group. Hyaline cartilage was retained in the subchondral bone of the MCP and the primary spongiosa, more so in the MCD- positive puppies than in MCD-negative puppies of both groups, and even more so in supplemented puppies than in control puppies (Figure 3, 4). Costal growth plates were more irregular in supplemented puppies than in control puppies, as evidenced by a difference in %sd of the growth plate thickness and by the less well organized columns of proliferating and hypertrophic chondrocytes (Table 3, Figure 4). The ratio between the width of the hypertrophic zone and the proliferation zone of the costal growth plate did not differ significantly between puppies of the control and supplemented groups (Table 3).

In both the MCD-negative and the MCD-positive puppies, collagen X staining was present 9 in the calcifying zone and in the non-mineralized cartilage matrix of the MCP (Figure 2), but significantly more stained surface area was demonstrated in the vitamin D-supplemented group compared to the control group, both in MCD-positive (p=0.039) as in MCD-negative (p=0.003) puppies (Figure 5, Table 3).

Discussion

We demonstrated in a previous study that endochondral ossification of the developing MCP is disturbed in MCD-positive growing Labrador retrievers (Lau et al. 2013b). During ossification, mineralization of the cartilaginous matrix is an important step in the replacement of the cartilage template by bone. The delayed maturation of the cartilage of the developing MCP results in a thicker cartilage layer as in dogs with MCD, can be caused by a deficiency in vitamin D or osteochondrosis (Thompson 2007). In addition, the presence of retained cartilage in the developing MCP makes the MCP more vulnerable to microtrauma and / or susceptible to relative mechanical overloading. Delayed skeletal maturation in large-breed dogs with rapidly increasing body weight and activity may cause mechanical damage to vulnerable cartilage (Olsson and Reiland 1978). Bearing in mind the role of vitamin D in skeletal development and mineralization (Hazewinkel and Tryfonidou 2002), we hypothesized that vitamin D supplementation of Labradors at risk of MCD would augment mineralization of the developing MCP at a young age and thus prevent the development of MCD.

To investigate this, we used Labrador retriever puppies at risk of developing MCD – their parents were a MCD-affected dam and a MCD-affected sire whose offspring were known to develop MCD (Lau et al. 2013b). Successful vitamin D supplementation was reflected by the significantly higher plasma levels of both 25-vitamin D and 1,25-vitamin D in the supplemented group compared with the control group. Although dietary vitamin D intake

140 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors was 43-fold increased, Ca and P plasma concentrations remained unchanged, which suggests that there was no vitamin D intoxication. This is in accordance with previous studies of young dogs raised on food with an increased vitamin D content, in which 24-hydroxylase activity was upregulated so that excess 25-vitamin D and 1,25-vitamin D was catabolized to prevent intoxication (Tryfonidou et al. 2003b).

Vitamin D influences skeletal development by stimulating the terminal differentiation of chondrocytes, matrix remodeling, and bone mineralization (Hazewinkel and Tryfonidou 2002). However, we found that vitamin D supplementation did not normalize endochondral ossification in the MCP area or prevent the development of MCD in Labrador puppies. Histological examination of the costal growth plates of the vitamin D-supplemented puppies showed the growth plates to be even more irregular with moderately disorganized chondrocyte columns than in the control puppies. Moreover, there appeared to be more retained cartilage in the primary spongiosa of the growth plate in the supplemented puppies than in the control puppies, which, taken together, is indicative of disturbed endochondral ossification. In line with these observations, the area within the MCP that stained positive for collagen type X was larger in supplemented puppies 9 than in control puppies, regardless of MCD status. On the basis of these observations, vitamin D supplementation would appear to predispose dogs to generalized disorders of endochondral ossification and hence to the development of orthopedic diseases. This is consistent with an earlier report in which growing Great Danes given a vitamin D-supplemented diet developed more severe disorders of endochondral ossification than Great Danes given a control diet (Tryfonidou et al. 2003a).

This study confirms that MCD is initiated within the deeper layers of the developing MCP rather than in the superficial layers of the articular cartilage (Lau et al. 2013b). After physiological regression of the cartilage canals, the persistence of hyaline cartilage within the developing MCP was demonstrated. In both groups, the prolonged presence of the unmineralized cartilage template in the MCP could predispose the MCP to fissure formation, but we cannot exclude that microtrauma precedes chondrocyte cell death. Whether mechanical loading is responsible for the disturbed mineralization of the cartilage template within the MCP remains to be established. MCD shares morphological similarities with osteochondrosis, a disease often seen at various skeletal sites in rapidly growing animals (Olsson and Reiland 1978, Lau et al. 2013b).

Conclusions

MCD is initiated in the deeper layers of the developing MCP and is characterized by delayed endochondral ossification and retention of cartilage within subchondral bone. While vitamin D supplementation did not accelerate chondrocyte maturation or matrix mineralisation of the cartilaginous MCP template in Labrador retriever puppies at risk of MCD development, it did disturb endochondral ossification, as evidenced by irregular growth plates and retained cartilage within the primary spongiosa of the costal growth plate.

141 Part II

Conflict of interest statement

None of the authors of this paper has a financial or personal relationship with other people or organizations that could inappropriately influence or bias the content of the paper.

Acknowledgments

The authors would like to acknowledge Dr. J. de Gier, Mrs. I. Maitimu-Smeele, Mr. M.R. Molenaar, Mrs. H.M. van Gils, Mr. H.G.H. van Engelen, and Mrs. I.I.M. van Duiven for their technical assistance. Dr. H.C.M. Heuven is acknowledged for his advises in the statistical analysis.

Appendix A: Supplementary material

Supplementary data associated with this article can be found, in the online version, 9 at doi: .

142 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

References • Howerth, E.W., 1983. Fatal soft tissue calcification in suckling puppies. Journal of • Corbee R.J., Tryfonidou M.A., Beckers I.P., the South African Veterinary Association 54, Hazewinkel H.A.W., 2012. Comparison of 21-24. the composition and the use of puppy milk • Lavrijsen, I.C., Heuven, H.C., Voorhout, G., Meij, replacers in German Shepherd puppies in the B.P., Theyse, L.F., Leegwater, P.A., Hazewinkel, Netherlands. Journal of Animal Physiology H.A., 2012. Phenotypic and genetic evaluation and Animal Nutrition 96, 395-402. of elbow dysplasia in Dutch Labrador • Danielson, K.C., Fitzpatrick, N., Muir, P., Manley, retrievers, Golden retrievers, and Bernese P.A., 2006. Histomorphometry of fragmented Mountain Dogs. The Veterinary Journal 193, medial coronoid process in dogs: a 486-492. comparison of affected and normal coronoid • Lau, S.F., Wolschrijn, C.F., Hazewinkel, H.A.W., processes. Veterinary Surgery 35, 501-509. Siebelt, M., Voorhout, G., 2013a. The early • Davidson, P.T., Bullock-Saxton, J., Lisle, A., development of medial coronoid disease in 2008. Anthropometric measurements of the growing Labrador retrievers: Radiographic, scapula, humerus, radius and ulna in Labrador computed tomographic, necropsy and dogs with and without elbow dysplasia. micro-computed tomographic findings. The 9 Australian Veterinary Journal 86, 425-428. Veterinary Journal 197, 724–730. • Fitzpatrick, N., Smith, T.J., Evans, R.B., Yeadon, • Lau, S.F., Hazewinkel, H.A.W., Grinwis, R., 2009. Radiographic and arthroscopic G.C.M., Wolschrijn, C.F., Siebelt, M., Vernooij, findings in the elbow joints of 263 dogs with J.C.M., Voorhout, G., Tryfonidou, M.A., medial coronoid disease. Veterinary Surgery 2013b. Delayed endochondral ossification 38, 213-223. in early medial coronoid disease (MCD): A • Gannon, J.M., Walker, G., Fischer, M., morphological and immunohistochemical Carpenter, R., Thompson, R.C. Jr., Oegema, T.R. evaluation in growing Labrador retrievers. The Jr., 1991. Localization of type X collagen in Veterinary Journal 197, 731–738. canine growth plate and adult canine articular • Olsson, S.E., Reiland, S., 1978. The nature cartilage. Journal of Orthopaedic Research 9, of osteochondrosis in animals. Summary 485-494. and conclusions with comparative aspects • Hazewinkel, H.A.W, Tryfonidou, M.A., 2002. on osteochondritis dissecans in man. Acta Vitamin D3 metabolism in dogs. Molecular Radiologica Supplement 358, 299-306. and Cellular Endocrinology 197, 23-33. • Sallander, M.H., Hedhammar, A., Trogen, • House, M.R., Marino, D.J., Lesser, M.L., 2009. M.E., 2006. Diet, exercise, and weight as risk Effect of limb position on elbow congruity factors in hip dysplasia and elbow arthrosis in with CT evaluation. Veterinary Surgery 38, Labrador retrievers. Journal of Nutrition 136, 154-160. 2050S-2052S. • How, K.L., Hazewinkel, H.A.W., Mol, J.A. 1994. • Testerink, N., Ajat, M., Houweling, M., Dietary vitamin D dependence of cat and Brouwers, J.F., Pully, V.V., van Manen, H., dog due to inadequate cutaneous synthesis Otto, C., Helms, J. & Vaandrager, A.B., of vitamin D. General and comparative 2012. Replacement of retinyl esters by endocrinology 96, 12-18. polyunsaturated triacylglycerol species in lipid droplets of hepatic stellate cells during activation, PLoS ONE, 7, e34945.

143 Part II

• Thompson, K., 2007. Bones and Joints. In: Jubb, Kennedy & Plamer’s Pathology of Domestic Animals (5th Ed), Editor: M. Grant Maxie, pp140. • Tryfonidou, M.A., Stevenhagen, J.J., Bemd, G.J.C.M. van den, Oosterlaken-Dijksterhuis, M.A., Deluca, H.F., Mol, J.A., Brom, W.E. van den, Van Leeuwen, J.P.T.M., Hazewinkel, H.A.W., 2002. Moderate cholecalciferol supplementation depresses intestinal calcium absorption in growing dogs. Journal of Nutrition 132, 2644-2650. • Tryfonidou, M.A., Holl, M.S., Stevenhagen, J.J., Buurman, C.J., Deluca, H.F., Oosterlaken- Dijksterhuis, M.A., Brom, W.E. van den, Leeuwen, J.P.T.M. van, Hazewinkel, H.A.W., 9 2003a. Dietary 135-fold cholecalciferol supplementation severely disturbs the endochondral ossification in growing dogs. Domestic Animal Endocrinology 24, 265-85. • Tryfonidou, M.A., Holl, M.S., Vastenburg, M., Oosterlaken-Dijksterhuis, M.A., Birkenhager- Frenkel, D.H., Brom, W.E. van den, Hazewinkel, H.A.W., 2003b. Hormonal regulation of calcium homeostasis in two breeds of dogs during growth at different rates. Journal of Animal Science 81, 1568-1580. • Van Ryssen, B., Bree, H. van, 1997. Arthroscopic findings in 100 dogs with elbow lameness. Veterinary Record 140, 360-362. • Voorhout, G., Hazewinkel, H. A., 1987. A radiographic study on the development of the antebrachium in Great Dane pups on different calcium intakes. Veterinary Radiology 28, 152-157.

144 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

Appendix A: Median stage of development of secondary ossification centers of the proximal ulnar apophysis, the medial humeral epicondyle and the ulnar styloid process (Voorhout and Hazewinkel 1987)

145 Part II

Development of secondary ossification centers (SOCs) in the elbow joint and distal ulna of puppies with (+) or without (-) medial coronoid disease from the control group (C) and the vitamin D-supplemented group (vitD)

C - C + vitD- vitD+ C - C + vitD- vitD+ C - C + vitD- vitD+ 9 weeks of age 12 weeks of age 16 weeks of age Olecranal apophysis Irregular shape Round 4 5 Round edges 3 2 3 2 1 1 1 Complete 6 6 3 1 7 7 3 2 ossification Distal Ulnar Metaphysis Flattened or 9 rounded Cartilage 6 7 3 2 cone Irregular 1 7 7 3 2 structure Normal shape 7 7 3 2 Medial Humeral epicondyle Round 4 7 1 2 Round edges 3 2 Complete 7 7 3 2 7 7 3 2 ossification

146 Dietary vitamin D supplementation during early growth does not protect against medial coronoid disease in Labradors

147 148 Chapter 10

Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism

Ronald J. Corbee Marianna A. Tryfonidou Björn P. Meij Hans S. Kooistra Herman A.W. Hazewinkel

Domestic Animal Endocrinology 2012; Volume 42, Issue 1, Pages 43-49

149 Part II

Abstract

Growth hormone (GH) and insuline like growth factor-1 (IGF-I) are important regulators of vitamin D metabolism in several species. Hypercortisolism and treatment with glucocorticoids in humans are associated with decreased 1,25-dihydroxy-vitamin D levels. Hypophysectomy is an effective therapy in dogs with pituitary- dependent hypercortisolism, resulting in lower GH and IGF-I blood levels and possibly 1,25-dihydroxy-vitamin D deficiency affecting calcium metabolism with eventually hyperparathyroidism. The aim of this study was to determine whether dogs with pituitary- dependent hypercortisolism need vitamin D supplementation before and/or after hypophysectomy. To this end we determined plasma concentrations of GH, IGF-I, parathyroid hormone, calcium, phosphorus and vitamin D metabolites in 12 dogs with pituitary-dependent hypercortisolism, before and 8 weeks after hypophysectomy, and in 12 age- and breed-matched, control dogs. We concluded that dogs with pituitary-dependent hypercortisolism 10 have lower GH levels when compared to control dogs, but do not have an altered vitamin D metabolism. The GH and IGF-I levels decreased after hypophysectomy, but all other parameters did not differ significantly between groups. After hypophysectomy down regulation of the GH-IGF-I-axis has no influence on 1,25-dihydroxyvitamin D levels nor on 24,25-dihydroxy-vitamin D levels. There is no need for vitamin D supplementation in dogs with pituitary- dependent hypercortisolism before or after hypophysectomy.

150 Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism

Introduction

Calcitriol (1,25-dihydroxy-vitamin D) is considered as one of the calciotropic hormones involved with calcium homeostasis. Its synthesis in the renal tubuli is under the influence of both other calciotropic hormones (i.e. parathyroid hormone (PTH) and calcitonin (CT)), as well as other factors. The main biological functions of 1,25-dihydroxy-vitamin D are stimulation of Ca and P intestinal absorption, renal reabsorption and osteoclasia as well as mineralization of newly formed osteoid and growth plate cartilage, the latter only in growing animals (Hazewinkel and Tryfonidou 2002). For vitamin D to become biologically effective it has to be activated by 25-hydroxylation and 1-α-hydroxylation into 25-hydroxy-vitamin D and 1,25-dihydroxy-vitamin D, respectively. It is deactivated by 24-hydroxylation into either 24,25-dihydroxy-vitamin D or 1,24,25-trihydroxy-vitamin D. Because dogs are unable to synthesize vitamin D in the skin under influence of UV B-light, dogs rely solely on dietary vitamin D intake (How et al. 1994). After dietary intake, vitamin D is hydroxylated into 25-hydroxy-vitamin D in the liver. The 25-hydroxy-vitamin D concentration of plasma thus reflects dietary vitamin D intake. The 1-α-hydroxylation in the kidney is stimulated by several factors, including plasma levels of Ca, P, PTH, growth hormone (GH), insulin-like growth factor-1 (IGF-I) and inhibited by 1,25-dihydroxyvitamin D and CT (Bianda et al. 1997, Halloran and Spencer 1988, Hazewinkel and 10 Tryfonidou 2002, Menaa et al. 1995, Nesbitt and Drezner 1993). Plasma 1,25-dihydroxyvitamin D levels are increased in young dogs after exogenously administered GH, which is attributed to an increased production of 1,25-dihydroxy-vitamin D mediated by IGF-I (Nap et al. 1993). Large breed dogs younger than 6 months of age with physiologically elevated GH and IGF-I levels have elevated 1,25-dihydroxy-vitamin D levels compared to small breed dogs of the same age The latter is attributed to a decreased clearance of 1,25-dihydroxy-vitamin D thereby keeping the levels in pace with the high demands of the rapidly-growing skeleton in these dogs with juvenile gigantism (Nap et al. 1993, Tryfonidou et al. 2003). From these studies it can be concluded that GH (directly) and/or IGF-I (indirectly) enhance 1-α-hydroxylation of 25-hydroxy-vitamin D.

Hypercortisolism results in an increased release of endogenous corticosteroids. The glucocorticoid induced osteoporosis in man with hypercortisolism was first described by Harvey Cushing (1932) and is the most common cause of secondary osteoporosis. Osteoporotic fractures are seen in 19-50% of human patients with hypercortisolism (Mancini et al. 2004, Ross and Linch 1982). Hypercortisolism influences bone metabolism in different ways; i.e., inhibition of osteoblastogenesis, stimulating activation and differentiation of osteoclasts, enhanced renal calcium excretion and decreased enteral absorption of calcium (Schäffler et al. 2009). In humans, corticosteroid treatment decreases GH, PTH and 1,25-dihydroxy-vitamin D plasma levels in a dose-dependent manner, resulting in osteopenia (Klaus et al. 2000, Suzuki et al. 1983). Prednisone-induced osteopenia is also demonstrated in dogs (Quarles et al 1992). Although osteopenia does not appear to be an important clinical feature in dogs with hypercortisolism, it can not be excluded that endogenous corticoids induce altered PTH levels and thereby vitamin D metabolism.

151 Part II

Hypophysectomy is an effective treatment in dogs suffering from pituitary-dependent hypercortisolism (Hanson et al. 2005, 2007, Hara et al. 2003, Meij 1998). After hypophysectomy the release of pituitary hormones including thyroid stimulating hormone, follicle stimulating hormone, luteinizing hormone, prolactin, growth hormone (GH), melanocyte stimulating hormone, adrenocorticotropic hormone (ACTH), and arginine vasopressine is impaired (Meij et al. 1997a, Meij et al. 1997b). In GH deficiency, the target hormone of GH, i.e., insulin-growth factor 1 (IGF-I), is hypothesized to be impaired as well. GH deficiency has been shown to cause 1,25-dihydroxy-vitamin D deficiency in children and hypophysectomized rats (De Boer et al. 1998), (Bruns et al. 1983, Chaudhry et al. 2009). A decrease in 1,25-dihydroxy-vitamin D decreases active intestinal calcium absorption and may cause hyperparathyroidism, with less renal excretion of calcium and increased osteoclastic osteoporosis. Hyperparathyroidism has been associated in dogs with a decreased quality of life, appetite, activity, strength and life span (Nagode et al. 1996).

The aim of this study is to determine whether dogs with hypercortisolism develop an altered PTH-vitamin D metabolism, and whether hypophysectomy leads to lower GH and IGF-I blood levels resulting in 1,25-dihydroxy-vitamin D deficiency and 10 hyperparathyroidism. The clinical value is to determine whether dogs with pituitary- dependent hypercortisolism need vitamin D supplementation before and/or after hypophysectomy.

Materials and methods Animals The study was approved by the Ethical Committee of the Faculty of Veterinary Medicine of Utrecht University. Permission for blood sampling was obtained from the owners of the dogs by informed consent. Inclusion criteria were confirmed diagnosis of pituitary- dependent hypercortisolism, no concurrent diseases, and hypophysectomy treatment followed by remission 8 weeks post-treatment. Twelve client-owned dogs diagnosed with pituitary-dependent hypercortisolism and treated by hypophysectomy were included in this study. Twelve healthy client-owned, age- and breed-matched dogs served as control dogs.

Diagnosis of pituitary-dependent hypercortisolism and hypophysectomy Diagnosis of pituitary dependent-hypercortisolism was made by clinical examination, results of hematology and blood chemistry, and results of the urine high-dose dexamethasone suppression test (i.e. 0.1mg/kg IV) (Kooistra and Galac 2010). Hypercortisolism was diagnosed when urine corticoid:creatinine ratios (UCCRs) were elevated in the first two urine samples (reference: <8.3 x 10-6 (van Vonderen et al. 1998)) and pituitary-dependent hypercortisolism was diagnosed when UCCR in the third sample was suppressed by more than 50% in comparison with the mean of the first two samples (Kooistra and Galac 2010). The diagnosis was further supported by measurement of plasma ACTH, abdominal ultrasonography of the adrenal glands and pituitary imaging

152 Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism

with computed tomography to determine tumor size and location. Hypophysectomy was performed using a microsurgical technique as described previously (Meij et al. 1997). After hypophysectomy, the dogs were substituted lifelong with levothyroxine and cortisone acetate, and desmopressine for 2 weeks or longer if necessary in case polyuria persisted due to central diabetes insipidus (Meij et al. 1998).

Blood sampling Before surgery blood was drawn from the dogs with pituitary-dependent hypercortisolism and from the control dogs; the blood was collected between 8 and 10 AM, in duplo (10 min apart) in ice-chilled EDTA-coated tubes, centrifuged at 4ºC and plasma was immediately frozen at -70ºC for further analysis; 1.5 ml blood was collected in ice-chilled heparin containing tubes, centrifuged and plasma stored at -70ºC for further analysis; and 2 ml blood was collected in ice-chilled serum tubes; after 30 minutes the blood was centrifuged and serum was stored at -70ºC for further analysis. Eight weeks after hypophysectomy blood sampling was repeated and processed in similar ways.

CBC and blood chemistry A CBC was performed as well as a biochemistry profile. Ca, P, total protein, albumin, BUN, creatinine, glucose, sodium, potassium, osmolarity, alkaline phosphatase (+65) and bile 10 acids were measured according to standard procedures (Beckman Industries Inc., Brea, USA).

Urinalysis and analysis of urinary corticoid:creatinine ratio’s Urinalysis was performed. Urine samples were collected in the morning by the owner at home, free of stress, on 2 consecutive days. After the second urine collection the owner administered three times 0.1mg/kg body weight dexamethasone orally at 8-hour intervals as described before (Kooistra and Galac 2010). The next morning a third morning urine sample was collected. Urinary corticoid concentrations were measured by radioimmunoassay (RIA) as described before (Rijnberk et al. 1988). The intra- and inter- assay CVs were 6 and 8% respectively, and the sensitivity was 0.36 ng/ml. The urinary corticoid concentration was related to the urinary creatinine concentration measured by the Jaffé initial rate kinetic method (Synchron CX Systems; Beckman Coulter) by calculating the UCCR.

Vitamin D metabolites, PTH, GH, IGF-I, and T4 Determination techniques of the plasma vitamin D metabolites levels have been described earlier and validated for the dog (Tryfonidou et al. 2002). Briefly, 25-hydroxy- vitamin D and 24,25-dihydroxy-vitamin D were quantitatively determined by a modified RIA (DiaSorin, Stillwater, Minnesota, USA) after extraction and separation by solid phase extraction (NH2 cartridge; Bakerbond spe Amino Disposable Extraction Columns, J.T. Baker, Philipsburg, USA). The intra- and inter-assay CV for 25-hydroxy-vitamin D were 15.2 and 6.1%, respectively. The intra- and inter-assay CV for 24,25-dihydroxy-vitamin D were 10.1 and 8.5%, respectively. 1,25-dyhydroxy-vitamin D was quantitatively determined by

153 Part II

a radio receptor assay based on the method described by (Hollis 1986, Reinhardt et al. 1984) after extraction with acetonitrile followed by a two step solid phase extraction (C18 and Silicagel cartridge; Waters Chromatography B.V. Etten Leur, The Netherlands) with an intra- and inter-assay CV of 5.7 and 6.6%, respectively. PTH was measured using an immunoradiometric assay for intact PTH (iPTH, Nichols Institute, San Juan Capistrano, CA, USA). The intra- and inter-assay CV were 3.4 and 5.6%, respectively. GH was measured by a homologous RIA as described before. The intra- and inter-assay CV were 3.8 and 7.2%, respectively. Total IGF-I levels were measured by a heterologous RIA as described previously with intra- and inter-assay CV of 4.7 and 15.6%, respectively (Tryfonidou et al. 2002).

Total T4 (TT4) was determined using a homologous solid-phase chemiluminescence enzyme immunoassay in accordance with the instructions of the manufacturer and validated for use in the dog (Bruner et al. 1998). The intra-assay coefficients of variation (CV) were 13.8 and 8.2% at TT4 concentrations of 8 and 25 nmol/l, respectively. The interassay CV was 8.5% at TT4 concentration of 21 nmol/l. The sensitivity was a TT4 concentration of 2 nmol/l.

10 Statistical analysis Statistical analyses were performed using SPSS 16.0 for Windows (SPSS Inc, Chicago, USA). Data were tested for normal distribution by a Kolmogorov-Smirnov test. All data were normally distributed accept for GH, IGF-I and PTH. Differences between the test group and the control group were analyzed by the two-tailed Student’s t-test. A Mann- Whitney-U-test was performed for non-parametric values and a one-way ANOVA was used to compare the means between groups. Differences between the test group before and after hypophysectomy were analyzed by a paired t-test. A Wilcoxon signed rank test was performed for non-parametric values. Values were considered to be significant when p<0.05. Results are presented as mean ± SEM.

Results

The patient group included 5 crossbreds, 2 Beagles, 1 Golden Retriever, 1 German pointing dog, 1 Dachshund, 1 French Bulldog and 1 miniature Poodle. The age of the patient group was 10.9±1.8 years. All were female neutered dogs. The control group included 3 crossbreds, 2 Beagles, 1 Golden Retriever, 1 German pointing dog, 1 Dachshund, 1 French Bulldog, 1 miniature Poodle, 1 West Highland White terrier and 1 Border terrier. The age of the control group was 9.9±2.3 years and did not differ from the test group (p= 0.273). All controls were also female neutered dogs.

The eleven dogs in the patient group had UCCRs that were inhibited by high dosage of dexamethasone (i.e. 0.1mg/kg IV) as is shown in table 1. The 12th dog included in this study appeared to have a pituitary tumor on the CT scan. Ten dogs had complete remission of the pituitary tumor after hypophysectomy, while in 2 dogs there was recurrence 2 and

154 Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism

Table 1: Urinary corticoid:creatinine ratios (UCCR) and % dexamethasone suppression in dogs with pituitary- dependent hypercortisolism

Dog Breed Gender Age UCCRa % (years) 1 2 3 suppressionw 1 Mongrel FN 10.1 41 34 5 86.7 2 Beagle FN 11.7 152 136 23 84.0 3 Golden retriever FN 10.8 140 167 57 62.9 4 Mongrel FN 9.3 17 14 3,6 76.8 5 Mongrel FN 11 37 35 1.4 96.1 6 German pointer FN 11.3 109 92 17 83.1 7 Mongrel FN 12.2 23 26 5.2 78.8 8 Dachshund FN 11.9 85 87 26 69.8 9 Mongrel FN 12.1 36 76 173 -208.9 10 French bulldog FN 7 308 242 43 84.4 11 Toy poodle FN 10.4 88 101 7.6 92.0 12 Beagle FN 6.8 46 51 7 85.6 a UCCRs measured in morning urine samples collected at home on day 1,2, and 3 respectively. After urine collection on day 2, three times 0.1mg/kg body weight dexamethasone was administered orally at 8-hour intervals. 10 b percentage of suppression of UCCR3 in comparison with the mean of UCCR 1 and UCCR2 FN = female neutered

Table 2: Mean plasma concentrations (±SD) of Ca, p, albumin, GH, IGF-I, PTH, 25-vit D, 24,25-vit D, and 1,25-vit D in dogs with pituitary-dependent hypercortisolism (n=12) before and after hypophysectomy (HX) and in breed- and age-matched control dogs (n=12)

Before HX After HX Control p values 1 2 3 1-3 1-2 2-3 Ca (mmol/L) 2.7±0.19 2.7±0.21 2.8±0.15 0.102 0.774 0.266 p (mmol/L) 1.6±0.34 1.4±0.25 1.2±0.27 0.002* 0.132 0.066 Albumin (g/L) 34.6±315 32.8±3.61 33.2±4.42 0.225 0.181 0.947 GH (ng/mL) 1.5±1.26 0.8±0.33 2.7±2.10 0.003* 0.043* 0.003* IGF-I (ng/mL) 87.8±42.53 53.8±10.23 109.9±51.37 0.163 0.015* 0.001* PTH (pg/mL) 79.3±39.18 83.75±42.6 76.0±63.86 0.205 0.397 0.464 25-vit D (ng/mL) 38.1±11.64 39.0±11.10 38.5±14.47 0.472 0.426 0.926 24,25-vit D (ng/mL) 30.6±11.84 30.8±11.83 40.0±17.01 0.131 0.480 0.141 1,25-vit D (pg/mL) 77.8±20.26 78.1±20.61 76.2±15.3 0.414 0.482 0.792

GH= growth hormone, IGF-I= insulin-like growth factor 1, PTH= parathyroid hormone, 25-vitD = 25-hydroxy-vitamin D; 24,25-vit D= 24,25-dihydroxy-vitamin D; 1,25-vitD= 1,25-dihydroxy-vitamin D *p<0.05, (1-3 and 2-3 Mann-Whitney-U-test; 1-2 Wilcoxon signed rank test)

155 Part II

4 years after surgery respectively, which was beyond the 6-8 week follow up period of our study. These dogs were thereafter treated with o,p’-DDD (Mitotane, Lysodren® 50- 75mg/kg per day per os, divided over 3-4 times daily during 5 days followed by 20 times every other day. Thereafter one day per 10-14 days in the same dosage). Pathology of the pituitary tumors all revealed to be pituitary adenomas. None of the dogs had a need for additional AVP (Minrin®) supplementation after surgery. GH levels were significantly lower in dogs with pituitary-dependent hypercortisolism after hypophysectomy (0.77±0.33ng/ mL) compared to pre-treatment levels (1.49±1.26ng/mL) (p=0.043) and compared to control dogs (2.73±2.10ng/mL) (p=0.003) (Fig 1). IGF-I levels were significantly lower in dogs with pituitary-dependent hypercortisolism after hypophysectomy (53.75±10.23ng/ mL) compared to pre-treatment levels (87.75±42.53ng/mL) (p=0.015) and compared to control dogs (109.92±51.37ng/mL) (p=0.001) (Fig 1). GH levels were already significantly lower in dogs with pituitary-dependent hypercortisolism (1.59±0.34ng/mL) compared to control dogs (1.16±0.27ng/mL) (p=0.003). Phosphate levels were significantly higher in dogs with pituitary-dependent hypercortisolism (1.49±1.26mmol/L) compared to control dogs (2.73±2.10mmol/L) (p=0.002). Blood levels of Ca, albumin, PTH, 25-hydroxy- vitamin D, 24,25-dihydroxyvitamin D and 1,25-dihydroxy-vitamin D did not differ significantly between dogs with pituitary-dependent hypercortisolism before and after 10 hypophysectomy and control dogs (Table 2).

Discussion

We hypothesized that dogs with hypercortisolism have an altered vitamin D metabolism resulting in osteopenia. Prednisone-induced osteopenia is demonstrated in dogs (i.e. dosage 1.3mg/kg/day during a 29 week period) by a suppression of osteoblastic function and recruitment, unrelated to PTH and 1,25-dihydroxy-vitamin D (Quarles et al 1992). However, another study demonstrated a fall of 1,25-dihydroxy-vitamin D levels in dogs treated with prednisone (1.3mg/kg/day) (Korkor et al. 1985). This was also accompanied by a decrease in intestinal calcium absorption. In humans, the urinary calcium excretion increases under glucocorticoid treatment which is suggested to be responsible for osteoporosis development (Suzuki et al. 1983). In pigs glucocorticoid treatment was associated with a fall in 1,25-hydroxy vitamin D as well as a fall in calcium absorption even after 1,25-dihydroxy-vitamin D levels had reached their minimum plateau, suggesting that another factor also plays a role (Fox et al. 1985). Based on the findings of our study, dogs with pituitary-dependent hypercortisolism have similar blood levels of PTH and vitamin D metabolites compared to healthy, age- and breed-matched control dogs, demonstrating no evidence for altered vitamin D metabolism or osteopenia.

The dogs with pituitary-dependent hypercortisolism had lower plasma GH concentrations compared to control dogs. This can be explained by impaired somatic growth reduced GH release, and a blocked GH response to various GH stimuli as is also demonstrated in other studies (Peterson and Altszuler 1981, Bhatti et al. 2002). After hypophysectomy there was a further decrease of plasma GH levels, as well as decreased plasma IGF-I levels.

156 Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism

This further decrease in plasma GH levels can be explained by the removal of the pituitary gland. There is still local production of GH, thus the plasma GH levels will remain at a certain level. The decrease in IGF-I levels after hypophysectomy has, to our knowledge, not been described before. We conclude that the IGF-I plasma levels decreased due to decreased levels of GH.

Dogs in all groups had 25-hydroxy-vitamin D levels in the range of 14.5-61.4μg/mL, reflecting sufficient dietary vitamin D intake (Hazewinkel and Tryfonidou 2002). Factors of influence for the hydroxylation of 25-hydroxy-vitamin D into 1,25-dihydroxy-vitamin D include Ca, P, PTH, GH and IGF-I. It was hypothesized that dogs after hypophysectomy lost part of the stimulation to activate 1-α-hydroxylase, leading to a deficiency of 1,25-dihydroxy-vitamin D activity with decreased Ca absorption and hyperparathyroidism as a result. The GH levels and IGF-I levels were indeed lowered in hypophysectomized dogs, however this did not result in decreased 1,25-dihydroxy–vitamin D plasma levels. PTH values were also similar between groups. In addition, there were no differences in 24,25-dihydroxy-vitamin D plasma concentrations before and after hypophysectomy, reflecting that vitamin D metabolism in dogs before and after hypophysectomy is not altered (Hazewinkel and Tryfonidou 2002). The explanation may be found in the old age of the animals (mean age 11 years): dogs of those ages have extremely low calcium 10 requirement, i.e. 0.5 mmol/kg body weight per day, only to correct for endogenous losses (Meyer and Zentek 2005). This requirement is covered by 2g Ca per kg food, being 1/6th of the Ca content of most commercially available dog foods (NRC 2006). Young dogs have a higher demand for 1,25-dihydroxy-vitamin D compared to mature, healthy dogs and probably the influences of GH on vitamin D metabolism are only important in case of an extra demand for 1,25-dihydroxy-vitamin D. This is supported by other studies, demonstrating an effect of GH on vitamin D metabolism in animals with an increased need for 1,25-dihydroxy-vitamin D like in case of vitamin D deficient food, low calcium or low phosphorus food intake, idiopathic hypophosphatemia and young, growing animals (Gray 1987, Nesbitt and Drezner 1993).

Conclusions

In dogs with pituitary-dependent hypercortisolism, increased cortisol levels do not affect the calciotropic hormones PTH and 1,25-dihydroxy-vitamin D. Also, after hypophysectomy, the decrease in GH and IGF-I has no significant influence on the plasma concentrations of these calciotropic hormones. Therefore, there is no need for vitamin D supplementation in dogs with pituitary-dependent hypercortisolism, either before or after hypophysectomy.

157 Part II

References • Fox, J., Ross, R., Care, A.D., 1985. Effects of acute and chronic treatment with • Bhatti, S.F.M., De Vliegher, S.P., Van Ham, glucocorticoids on the intestinal absorption L., Kooistra, H.S., 2002. Effects of growth of calcium and phosphate and on plasma hormone-releasing peptides in healthy 1,25-dihydroxyvitamin D levels in pigs. dogs and in dogs with pituitary-dependent Clinical Science 69, 553-559. hyperadrenocorticism. Molecular and Cellular • Gray, R.W., 1987. Evidence that somatomedins Endocrinology 197, 97-103. mediate the effect of hypophosphatemia to • Bianda, T., Hussain, M.A., Glatz, Y., Bouillon, increase serum 1,25-dihydroxyvitamin D3 R., Froesch, E.R., Schmid, C., 1997. Effects of levels in rats. Endocrinology 121, 504-512. short-term insulin-like growth factor-I or • Halloran, B.P., Spencer, E.M., 1988. Dietary growth hormone treatment on bone turnover, phosphorus and 1,25-dihydroxyvitamin D renal phosphate reabsorption and 1,25 metabolism: Influence of insulin-like growth dihydroxyvitamin D3 production in healthy factor I. Endocrinology 123, 1225-1229. man. Journal of Internal Medicine 241, • Hanson, J.M., Van ‘t Hoofd, M.M., Voorhout, 143-150. G., Teske, E., Kooistra, H.S., Meij, B.P., 2005. • Bruner, J.M., Scott-Moncrieff, J.C.R., Williams, Efficacy of transsphenoidal hypophysectomy D.A., 1998. Effect of time of sample collection in treatment of dogs with pituitary- 10 on serum thyroid-stimulating hormone dependent hyperadrenocorticism. Journal of concentrations in euthyroid and hypothyroid Veterinary Internal Medicine 19, 687-694. dogs. Journal of the American Veterinary • Hanson, J.M., Teske, E., Voorhout, G., Galac, Medical Association 212, 1572-1575. S., Kooistra, H.S., Meij, B.P., 2007. Prognostic • Bruns, M.E.H., Vollmer, S.S., Bruns, D.E., factors for outcome after transsphenoidal Overpeck, J.G., 1983. Human growth hypophysectomy in dogs with pituitary- hormone increases intestinal vitamin dependent hyperadrenocorticism. Journal of D-dependent calcium-binding protein in Neurosurgery 107, 830-840. hypophysectomized rats. Endocrinology 113, • Hara, Y., Tagawa, M., Masuda, H., Sako, 1387-1392. T., Koyama, H., Orima, H., Nakamura, S., • Chaudhry, A.A., Castro-Magana, M., Aloia, J.F., Takahashi, K., Sanno, N., Teramoto, A., 2003. Yeh, J.K., 2009. Differential effects of growth Transsphenoidal hypophysectomy for hormone and alfa calcidol on trabecular and four dogs with pituitary ACTH-producing cortical bones in hypophysectomized rats. adenoma. Journal of Veterinary Medical Pediatric Research 65, 403-408. Science 65, 801-804. • De Boer, H., Blok, G.J., Popp-Snijders, C., Sips, • Hara, Y., Teshima, T., Taoda, T., Ishino, H., A., Lips, P., Van Der Veen, E., 1998. Intestinal Nezu, Y., Harada, Y., Yogo, T., Masuda, H., calcium absorption and bone metabolism Teramoto, A., Tagawa, M., 2010. Efficacy of in young adult men with childhood-onset transsphenoidal surgery on endocrinological growth hormone deficiency. Journal of Bone status and serum chemistry parameters and Mineral Research 13, 245-252. in dogs with Cushing’s disease. Journal of Veterinary Medical Science 72, 397-404. • Hazewinkel, H.A.W., Tryfonidou, M.A., 2002. Vitamin D3 metabolism in dogs. Molecular and Cellular Endocrinology 197, 23-33.

158 Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism

• Hollis, B.W., 1986. Assay of circulating • Meij, B.P., Mol, J.A., Van Den Ingh, T.S.G.A.M., 1,25-dihydroxyvitamin D involving a novel Bevers, M.M., Hazewinkel, H.A.W., Rijnberk, A., single-cartridge extraction and purification 1997b. Assessment of pituitary function after procedure. Clinical Chemistry 32, 2060-2063. transsphenoidal hypophysectomy in beagle • How, K.L., Hazewinkel, H.A.W., Mol, J.A., 1994. dogs. Domestic Animal Endocrinology 14, Dietary vitamin D dependence of cat and 81-97. dog due to inadequate cutaneous synthesis • Meij, B.P., Voorhout, G., Van Den Ingh, of vitamin D. General and comparative T.S.G.A.M., Hazewinkel, H.A.W., Teske, E., endocrinology 96, 12-18. Rijnberk, A., 1998. Results of transsphenoidal • Klaus, G., Jux, C., Fernandez, P., Rodriguez, hypophysectomy in 52 dogs with pituitary- J., Himmele, R., Mehls, O., 2000. Suppression dependent hyperadrenocorticism. Veterinary of growth plate chondrocyte proliferation Surgery 27, 246-261. by corticosteroids. Pediatric Nephrology 14, • Meij, B.P., Voorhout, G., Van Den Ingh, 612-615. T.S.G.A.M., Hazewinkel, H.A.W., Van ‘t Verlaat, • Kooistra, H.S., Galac, S., 2010. Recent J.W., 1997. Transsphenoidal hypophysectomy Advances in the Diagnosis of Cushing’s in beagle dogs: Evaluation of a microsurgical Syndrome in Dogs. Veterinary Clinics of North technique. Veterinary Surgery 26, 295-309. America - Small Animal Practice 40, 259-267. • Menaa, C., Vrtovsnik, F., Friedlander, G., • Korkor, A.B., Kuchibotla, J., Arrieh, M., Corvol, M., Garabedian, M., 1995. Insulin-like 10 1985. The effects of chronic prednisone growth factor I, a unique calcium-dependent administration on intestinal receptors stimulator of 1,25-dihydroxyvitamin D3 for 1,25-dihydroxyvitamin D3 in the dog. production. Studies in cultured mouse kidney Endocrinology 117, 2267-2273. cells. Journal of Biological Chemistry 270, • Mancini, T., Doga, M., Mazziotti, G., Giustina, 25461-25467. A., 2004. Cushing’s syndrome and bone. • Nagode, L.A., Chew, D.J., Podell, M., Pituitary 7, 249-252. 1996. Benefits of calcitriol therapy and • Meij, B., Voorhout, G., Rijnberk, A., 2002. serum phosphorus control in dogs and Progress in transsphenoidal hypophysectomy cats with chronic renal failure. Both are for treatment of pituitary-dependent essential to prevent or suppress toxic hyperadrenocorticism in dogs and cats. hyperparathyroidism. Veterinary Clinics of Molecular and Cellular Endocrinology 29, North America - Small Animal Practice 26, 89-96. 1293-1330. • Meij, B.P., Mol, J.A., Bevers, M.M., Rijnberk, • Nap, R.C., Mol, J.A., Hazewinkel, H.A.W., 1993. A., 1997a. Residual pituitary function Age-related plasma concentrations of growth after transsphenoidal hypophysectomy hormone (GH) and insulin-like growth factor in dogs with pituitary-dependent I(IGF-I) in Great Dane pups fed different hyperadrenocorticism. Journal of dietary levels of protein. Domestic Animal Endocrinology 155, 531-539. Endocrinology 10, 237-247. • Nesbitt, T., Drezner, M.K., 1993. Insulin- like growth factor-I regulation of renal 25-hydroxyvitamin D-1- hydroxylase activity. Endocrinology 132, 133-138.

159 Part II

• Peterson, M.E., Altszuler N., 1981. Suppression • Tryfonidou, M.A., Stevenhagen, J.J., Van Den of growth hormone secretion in spontaneous Bemd, G.J.C.M., Oosterlaken-Dijksterhuis, M.A., hyperadrenocorticism and its reversal after Deluca, H.F., Mol, J.A., Van Den Brom, W.E., Van treatment. American Journal of Veterinary Leeuwen, J.P.T.M., Hazewinkel, H.A.W., 2002. Research 42, 1881-1883. Moderate cholecalciferol supplementation • Reinhardt, T.A., Horst, R.L., Orf, J.W., depresses intestinal calcium absorption in Hollis, B.W., 1984. A microassay for growing dogs. Journal of Nutrition 132, 1,25-dihydroxyvitamin D not requiring 2644-2650. high performance liquid chromatography: • van Vonderen, I.K., Kooistra, H.S., Rijnberk, Application to clinical studies. Journal of A., 1998. Influence of veterinary care on the Clinical Endocrinology and Metabolism 58, urinary corticoid:creatinine ratio in dogs. 91-98. Journal of Veterinary Internal Medicine 12, • Rijnberk, A., van Wees, A., Mol, J.A., 1988. 431-435. Assessment of two tests for the diagnosis of canine hyperadrenocorticism. Veterinary Record 122, 178-180. • Ross, E.J., Linch, D.C., 1982. Cushing’s syndrome - killing disease: Discriminatory 10 value of signs and symptoms aiding early diagnosis. Lancet 2, 646-647. • Schäffler, A., Bala, M., Bollheimer, C., Schölmerich, J., 2009. Hormone-Induced Spontaneous Femoral Neck Fracture in a 28-Year-Old Female Patient. Medizinische Klinik 104, 244-248. • Suzuki, Y., Ichikawa, Y., Saito, E., Homma, M., 1983. Importance of increased urinary calcium excretion in the development of secondary hyperparathyroidism of patients under glucocorticoid therapy. Metabolism: Clinical and Experimental 32, 151-156. • Tryfonidou, M.A., Holl, M.S., Vastenburg, M., Oosterlaken-Dijksterhuis, M.A., Birkenhäger- Frenkel, D.H., Van Den Brom, W.E., Hazewinkel, H.A.W., 2003. Hormonal regulation of calcium homeostasis in two breeds of dogs during growth at different rates. Journal of Animal Science 81, 1568-1580.

160 Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism

161 162 Chapter 11

Ground reaction forces of walking cats; assessment and comparison with walking dogs

Ronald J. Corbee Huub Maas Arie Doornenbal Herman A.W. Hazewinkel

The Veterinary Journal; DOI 10.1016/j.tvjl.2014.07.001

163 Part II

Abstract

The first aim of this study was to assess the potential of force plate analysis for describing the stride cycle of the cat. The second aim of this study was to define differences in locomotion of cats from that of dogs based on force plate characteristics. Ground reaction forces of 24 healthy cats were measured and compared with ground reaction forces of 24 healthy dogs. The force-time wave-forms in cats, generated by force plate analysis, were consistent as reflected by an intraclass correlation coefficient for peak vertical force, peak propulsive force and peak braking force of 0.94-0.95, 0.85-0.89, and 0.89-0.90, respectively. Cats had a higher peak vertical force during the propulsion phase (3.89±0.19N/kg vs. 3.03±0.16N/kg), and a higher propulsive force (-1.08±0.13N/ kg) and its impulse (-0.18±0.03Ns/kg) of the hindlimbs compared to dogs (-0.87±0.13N/kg, and -0.14±0.02Ns/kg, respectively). In conclusion, force plate analysis is a valuable tool for assessment of locomotion in cats, because it can be applied in the clinical setting, and provides a non-invasive and objective measurement of locomotion characteristics with high repeatability in cats, as 11 well as additive information about kinetic characteristics. Differences in force-time wave-forms between cats and dogs can be explained by the more crouched position of cats compared to dogs during stance, and more compliant gait of cats. Another difference can be found in the wave-forms of the medio-lateral ground reaction forces, which can be explained by differences in paw supination-pronation.

164 Ground reaction forces of walking cats; assessment and comparison with walking dogs

Introduction

Orthopedic diseases in cats can cause lameness, but lameness is not an obvious symptom of the most common orthopedic disease in cats: osteoarthritis (OA). OA is a frequently encountered disorder in ageing cats with a prevalence of 22-72% in cats over 6 years of age (Hardie et al. 2002, Clarke et al. 2005, Godfrey 2005). The prevalence and severity of OA increase with age (Lascelles 2010, Slingerland et al. 2011). Abaxially, the elbows, hips, shoulders and tarsi are the most commonly affected joints (Slingerland et al. 2011). Axially the thoracic vertebrae Th4-10 are most frequently affected, but the lumbosacral vertebrae are most severely affected (Kranenburg et al. 2012). Early diagnosis and treatment of OA are important, because of the consequences of OA on quality of life (Bennett and Morton 2009, Corbee et al. 2013, Kranenburg et al. 2012, Guillot et al. 2013).

Diagnosis of orthopedic diseases can be challenging in cats, because they often do not tolerate a full orthopedic examination (Hardie et al. 2002, Zamprogno et al. 2010, Kranenburg et al. 2012). Therefore, less stressful methods are deemed necessary. The interpretation of orthopedic examination in cats is semi-subjective and largely depends on the observer (Lascelles et al. 2012). Determining the degree of lameness by the amount of osteophytes visible on radiographs is not reliable, since this does not correlate with lameness as assessed by force plate analysis (Suter et al. 1998), nor with pain, crepitus, or reduced range of joint motion upon orthopedic examination (Lascelles et al. 2012). More 11 often, subjective owner assessment, by the use of standardized questionnaires (client- specific outcome measures), is used to judge the presence of lameness or impairments during gait and other movements of cats (Slingerland et al. 2011, Benito et al. 2012, Benito et al. 2013a, Benito et al. 2013b). Owners find it difficult to score and cannot recognize pain or lameness due to OA, because stiffness, unkempt hair coat, and reluctance to jump, are considered by them as normal aspects of aging in cats (Kranenburg et al. 2012, Corbee et al. 2013). It is important for clinicians, as well as researchers, to monitor effect of treatment of OA in cats by more objective measures.

Several methods for gait analysis have been described in cats and dogs, such as accelerometry (Lascelles et al. 2010, Guillot et al. 2012, Mazurek et al. 2012, Guillot et al. 2013, Grand et al. 2013), imaging by high speed camera while having reflecting patches attached to the skin (Gillette et al. 2008), pressure-sensitive walkway (Lascelles et al. 2007, Rialland et al. 2012), and force plate analysis (Suter et al. 1998). Weighing scales and pressure-sensitive walkways provide an assessment of only the vertical ground reaction force (Fz). With a 3-dimensional force plate, also the cranio-caudal force (Fy) and the medio-lateral ground reaction force (Fx) exerted during the stance phase can be analyzed (Merkens 1987).

Force plate analysis has been used in veterinary research as a non-invasive, objective evaluation of lameness and/or evaluation of effect of treatment in dogs and horses (Theyse et al. 2000, Guedes et al. 2012, Hazewinkel et al. 2003, Kalis et al. 2012, Oomen

165 Part II

et al. 2012, Smolders et al. 2012, Spaak et al. 2012, Van der Peijl et al. 2012). The first study implementing force plate analysis for evaluation of locomotion in dogs with hip dysplasia was published by Dueland et al. (1977). After the publication of McLaughlin Jr. et al. (1991), force plate analysis became an objective standard in evaluation of canine locomotion. Clinical studies in cats with force plate analysis are scarce and generally limited to measurements of Fz (e.g. Moreau et al. 2013, Guillot et al. 2012, Grösslinger et al. 2007). The force-time wave-forms for all three directions may provide more insight when evaluating kinetic gait data (Fransz et al. 2013, Al-Nadaf et al. 2012). Most of the studies using force plate analysis in cats are related to comparative neurophysiology research and do not involve veterinary clinical applications (Prilutsky et al. 2011, Gregor et al. 2006, Suter et al. 1998). Because cats and dogs share similar orthopedic diseases, a comparison of ground reaction forces during over-ground walking is deemed a valuable first step in the assessment of force plate analysis as a diagnostic tool in cats.

The first aim of this study was to quantify 3-dimensional ground reaction forces during over-ground walking in the cat. The second aim of this study was to assess differences in ground reaction force patterns between cats and dogs.

Materials and methods Ethical statement 11 This study was conducted with permission of the ethical and welfare committee, as is required under Dutch legislation (NL DEC 2011.III.01.008).

Animal care and training Twenty-four healthy domestic Shorthair cats (12 males and 12 females; mean age 7 years, range 2-14 years; mean body weight 3.8kg, range 2.2-6.9kg) from the departmental cat colony, were trained to walk on a leash. Some cats easily walked on a leash, but most cats needed to get used to the environment, and to wear a cat harness. All cats got used to uninterrupted leash walking without accelerations in an average of five 30 minutes sessions within a 2-week training period. Prior to a training session, the cats were fasted and were allowed to adapt to the force plate room for 10 min. Kibbles and affection were used to encourage the cats to walk.

The data obtained in 24 healthy cats were compared with data of 24 healthy Labradors (eight males and 16 females; mean age 16 months, range 15-17 months; average body weight 26kg, range 21-32kg). The cats and dogs were considered healthy based on clinical and orthopedic examination by a board certified veterinary surgeon (HH) and normal urinalysis, complete blood count, and blood biochemistry. No abnormalities were observed on plain radiographs of the joints and vertebrae in all cats or dogs included in the study.

166 Ground reaction forces of walking cats; assessment and comparison with walking dogs

Fig. 1: Adaptation of the force plate for force plate analysis (FPA) in cats. An overlay was attached over the force plate. The middle plate was attached to the force plate by the use of two M8x40 bolts. The rest of the overlay does not touch the force plate and does not interfere with FPA 11

Table 1: Correlation coefficients and symmetry index of force plate analysis in cats

Fzmax Fymax Fymin LF RF LH RH LF RF LH RH LF RF LH RH ICC1 0.94 0.94 0.91 0.91 0.87 0.87 0.82 0.82 0.85 0.85 0.89 0.89 ICC2 0.94± 0.95± 0.94± 0.94± 0.89± 0.89± 0.85± 0.86± 0.89± 0.90± 0.89± 0.87± 0.03 0.02 0.02 0.02 0.04 0.05 0.04 0.06 0.04 0.04 0.05 0.06 L/R L/R L/R L/R L/R L/R SI 97±2 97±2 94±5 87±7 94±5 92±7

Data of 24 cats were analyzed. The interclass correlation coefficient (ICC1) was calculated for the group of cats. The intraclass correlation coefficient (ICC2) and the symmetry index were calculated for each individual cat and presented as mean ± standard deviation for 24 cats. Fz = Vertical force, Fy = Cranio-caudal force, Fx = Medio-lateral force, LF = Left frontlimb, RF = Right frontlimb, LH = Left hindlimb, RH = Right hindlimb, ICC1 = Interclass correlation coefficient, ICC2 = Intraclass correlation coefficient, SI = Symmetry index, L/R = Left to right symmetry

167 Part II

Table 2: Force plate data of cats. Explanation of data points are given in Fig. 2

Amplitude (N/kg) Position of stance time LF RF LH RH LF RF LH RH Fzmax1 5.69±0.37 5.70±0.38 4.83±0.49 4.76±0.48 31±1 32±1 25±0 25±0 Fzdip 5.45±0.38 5.48±0.37 3.84±0.20 3.87±0.21 52±1 51±1 55±1 56±1 Fzmax2 5.56±0.37 5.56±0.39 3.89±0.19 3.89±0.19 65±0 65±0 63±1 61±1 Iz 1.95±0.27 1.93±0.18 1.46±0.12 1.42±0.13

Fymax 1.44±0.21 1.44±0.23 0.87±0.19 0.81±0.21 23±1 24±1 17±0 17±0 t (Fy=0) 57±1 57±1 37±1 36±1 Fymin -1.02±0.18 -1.06±0.17 -1.08±0.13 -1.08±0.13 81±0 81±0 72±0 72±0 Iymax 0.22±0.04 0.22±0.04 0.08±0.02 0.07±0.02 Iymin -0.12±0.03 -0.12±0.03 -0.18±0.04 -0.18±0.03

Fxmax1 0.05±0.06 0.05±0.06 0.28±0.15 0.27±0.12 4±0 4±0 22±0 22±0 Fxdip 0.16±0.12 0.17±0.13 41±1 43±1 Fxmax2 0.39±0.12 0.39±0.14 0.28±0.15 0.27±0.12 30±0 30±0 69±0 69±0

Tangent ɑ 3.08±0.18 3.08±0.18 2.90±0.15 2.90±0.15 Tangent β 1.66±0.14 1.66±0.14 1.19±0.14 1.19±0.14 11 Tangent ϒ 1.28±0.15 1.28±0.15 1.19±0.16 1.19±0.14

N/kg, ground reaction force in N, per kg of bodyweight; L/R, left to right symmetry; % stance time, % of the total stance time at which the data points are reached; F, frontlimb; H; hindlimb; I, impulse in Ns per kg of bodyweight (Ns/ kg); t (Fy=0), time point at which there is no net propulsive force and no net braking force; tangent, tangent of the angle depicted in Fig. 2. Data of 10 trials were first averaged per animal and then analyzed across animals. The mean ± standard deviation (SD) presented is the mean ± SD for 24 cats. Bold numbers are significant different compared to dogs (p<0.05; ANOVA). The exact p-values are given in Table 3

Data collection The body weight of each cat was determined on an electronic scale (DIWAC VS150) and recorded immediately before force plate measurement. A quartz piezoelectric force plate (Kistler type 9261) with Kistler 9865B charge amplifiers, mounted flush in a walk way (5m for cats, 11m for dogs) was used. The walkway was enclosed by a fence to guide the animal over the force plate. The force platform area, standard 40cm long and 60cm wide for dogs, was decreased to 25cm long and 60cm wide by a firmly attached overlay plate for measurement of the cats (Fig. 1). Sampling rate was 100Hz. Amplifiers were connected to a computer that stored the signals, which corresponded with ground reaction forces in the vertical (Fz), craniocaudal (Fy), and mediolateral (Fx) directions.

Before data collection, equilibration and calibration of the force plate were performed, according to the manufacturer’s specifications. Data of all four legs were collected in

168 Ground reaction forces of walking cats; assessment and comparison with walking dogs

Table 3: Force plate data of dogs. Explanation of data points are given in Fig. 2

Amplitude (N/kg) Position of stance time LF RF LH RH LF RF LH RH Fzmax1 6.02±0.29 5.98±0.28 4.28±0.58 4.31±0.41 34±0 34±0 25±1 24±1 Fzdip 5.61±0.34 5.52±0.44 2.72±0.17a 2.79±0.19a 62±0a 62±0a 57±1 56±1 Fzmax2 5.69±0.36 5.63±0.42 3.01±0.17a 3.03±0.16a 70±0a 70±0a 76±0a 76±0a Iz 2.43±0.04a 2.41±0.05a 1.37±0.02 1.38±0.02

Fymax 1.30±0.11 1.31±0.14 0.78±0.19 0.80±0.15 23±1 23±1 17±0 17±0 t (Fy=0) 56±1 56±1 36±0 36±0 Fymin -1.06±0.14 -1.06±0.09 -0.86±0.11a -0.87±0.13a 82±0 82±0 82±0a 82±0a Iymax 0.18±0.03 0.19±0.03 0.07±0.02 0.07±0.02 Iymin -0.12±0.03 -0.12±0.02 -0.14±0.02a -0.14±0.02a

Fxmax1 0.23±0.15a 0.29±0.13a 0.34±0.16b 0.36±0.20c 15±1a 12±2a 25±1a 26±1a Fxmax2 0.57±0.22a 0.59±0.19a 0.20±0.09d 0.20±0.10e 35±1f 33±2g 80±0a 80±0a

Tangent ɑ 1.88±0.17a 1.88±0.16a 2.61±0.25 2.61±0.25 Tangent β 1.19±0.18h 1.19±0.19e 1.00±0.12 1.00±0.11 Tangent ϒ 1.38±0.22 1.38±0.20 1.00±0.11 1.00±0.12 11 N/kg, ground reaction force in N, per kg of bodyweight; % stance time, % of the total stance time at which the data points are reached; F, frontlimb; H; hindlimb; I, impulse in Newton second per kg of body weight (Ns/kg); t (Fy=0), time point at which there is no net propulsive force and no net braking force; tangent, tangent of the angle depicted in Fig. 2. Data of 10 trials were first averaged per animal and then analyzed across animals. The mean ± standard deviation (SD) presented is the mean ± SD for 24 dogs. Bold numbers are significant different compared to cats (P<0.05 ANOVA) a p<0.001, b p=0.044, c p=0.035, d p=0.028, e p=0.032, f p=0.001, g p=0.007, h p=0.029

10 trials in one session for each cat (n=24 cats). This was repeated after three weeks to determine intra-session and intersession variability, respectively. The trials collected in the first session were used for comparison with those obtained in dogs.

Average walking speed across trials and cats was 0.7 ± 0.1 m/s. It is important for the cats to walk with an uninterrupted gait at a constant speed to avoid differences due to a different gait type or duration of the stance phase (Halbertsma 1983).

Forward velocity was measured, using photoelectric switches and a ms timer (Hazewinkel et al. 2003). Starting and ending of measurement were automatically regulated when the cat passed the switches incorporated in the fence. The same person guided all cats on a leash over the force plate during all recordings without acceleration. Each pass across the platform was also evaluated by a single observer to confirm that the forelimb

169 Part II

was followed by the ipsilateral hindlimb in the same run, and that each foot contacted the force plate completely (Video 1). Trials were discarded for wrong walking speed, distracting head motions, irregularities in the gait, partial loading of the plate, or more than one foot simultaneously on the plate.

To obtain qualitative information about fore- and hindlimb movement patterns, we filmed three cats and three dogs during walking in the walkway using a high speed camera (Casio Exilim EX-F1 6 Megapixel 60fps Hi-speed) in both sagittal and frontal planes. A number of 10 trials per dog (n=3) and cat (n=3) were filmed. Typical examples are presented in Appendix A (Video 1-6).

Data analysis All ground reaction force data were normalized to body weight. Impulses were calculated by NI Lab view 8.2 software and presented in Newton x s per kg of body weight (Ns/kg body weight). Ground reaction force data was time-normalized with respect to stance time for direct comparison of force-time wave-forms within and between cats and dogs. The tangent of the peak angle of the initial force (Fz, Fy, Fx) is presented (as tangent α, β, γ, respectively) as a measure of the rate of increase of the force over time at the beginning of stance. All trials were averaged first within animals (i.e. average of 10 trials per individual animal) and then averaged across animals. Data are presented as mean 11 ± standard deviation of averaged force data across animals.

Characteristic points at the complete force-time wave-forms, i.e. Fzmax1, Fzdip, Fzmax 2, Fymax, Fymin, t Fy=0, Fxmax1, Fxdip, and Fxmax2 are indicated in Fig. 2, in addition to the angles α, β, and γ. The symmetry between left and right legs was calculated by calculating the ratio according to Mueller et al. (2007) and presented in Table 1.

Video data were synchronized to force plate data using switches in the fence that registered placement of mass on the force plate and coupled to a lamp to visualize this on the video (Appendix A, Video 1). The video data were recorded with 60 frames/s (fps), but displayed at 25 fps, and thus shown in slow motion (2.4 times slower).

Statistical analysis Statistical analyses were performed with the use of R-statistics (R i386 3.0.1). A Kolmogorov-Smirnov test was performed to test for normal distribution. All force plate data were normally distributed and tested by a paired t-test for different time points, and for differences between the left and right legs. One-way ANOVA was used to test for differences in selected parameters between cats and dogs. All these comparisons were based on average of the 10 steps per animal. p<0.05 was set as the level of significance. For demonstration of the reproducibility of the data, an intraclass correlation coefficient (ICC) was calculated. The ICC was considered high when it was above 0.90 (Bénard et al. 2010).

170 Ground reaction forces of walking cats; assessment and comparison with walking dogs

Fig. 2: Explanation of data points in the force plate analysis of cats. Fz, vertical force; Fy, cranio-caudal force; Fx, medio-lateral force; RF, right forelimb; RH, right hindlimb

Results Force plate data The left and right fore- and hindlimbs revealed symmetry in the amplitude for Fzmax, Fymax, Fymin of 87-97% (Table 1). Data for subsequent measures of the 10 steps revealed within day ICCs of 0.85-0.95 and between days ICCs of 0.82-0.94 (Table 1). Force plate data of cats and dogs are presented in Table 2 and Table 3, respectively. The force-time wave-forms of cats and dogs are presented in Fig. 3 and Fig. 4, respectively.

Cats - forelimbs The peak vertical force of the right forelimb (Fzmax1= 5.70±0.38N/kg) is reached at 32±1% of the stance phase, starting with a steep angle (tangent α = 3.08±0.18; Fig. 3). Fz remained almost the same during midstance (Fzdip = 5.48±0.37N/kg at 51±1% of the stance phase) and dropped after reaching a second peak (Fzmax2 = 5.56±0.39N/kg) at 65±0% of the stance phase. The vertical impulse (Iz) was 1.93±0.18Ns/kg.

171 Part II

Fig. 3: Force-time wave-forms of 24 normal cats, presented as mean ± standard deviation (SD). Data of 10 trials were first averaged per animal and then analyzed across animals. The mean ± SD presented is the mean ± SD from 24 cats. Fz, vertical force; Fy, cranio-caudal force; Fx, medio-lateral force; RF, right forelimb; RH, right hindlimb

172 Ground reaction forces of walking cats; assessment and comparison with walking dogs

Fig. 4: Force-time wave-forms of 24 normal dogs, presented as mean ± standard deviation (SD). Data of 10 trials were first averaged per animal and then analyzed across animals. The mean ± SD presented is the mean ± SD from 24 dogs. Fz, vertical force; Fy, cranio-caudal force; Fx, medio-lateral force

173 Part II

At the beginning of the stance phase of the forelimbs of cats, there was a brake force, which reached a maximum (Fymax = 1.44±0.23N/kg) at 24±1% of the stance phase, with a tangent β of 1.66±0.14 (Fig. 3). Maximum propulsive force (Fymin = -1.06±0.17N/kg) was found at 81±0% of the stance phase. The braking phase occupied 57±1% of the stance phase (t Fy=0), and thus the propulsive phase was 43±1% of the stance phase. The positive impulse (Iy+) was 0.22±0.04Ns/kg and the negative impulse (Iy-) was -0.12±0.03Ns/kg.

At the beginning of stance, Fx is directed laterally (- for the right leg and + for the left leg, Fig. 2). Fx returned to neutral position at 2±0% of the stance phase, and then reached a first maximum in a medial direction (Fxmax1= 0.05±0.06N/kg at 4±0% of the stance phase). During the braking phase it reached neutral position (at 10% of the stance phase), resulting in a medially directed (+) ground reaction force (Fxmax2 = 0.39±0.14N/kg at 30±0% of the stance phase), and a tangent γ of 1.28±0.15. At the end of the propulsive phase Fx returned to a neutral position (at 87% of the stance phase) and finally ends in the lateral direction.

Cats - hindlimbs The peak vertical force of the right hindlimb (Fzmax1 = 4.76±0.48N/kg) was found, with a steep angle (tangent α = 2.90±0.15) at 25±0% of the stance phase (Fig. 3). Then Fz 11 decreased and reached a plateau at approximately 48% of the stance phase (Fzdip = 3.87±0.21N/kg at 56±1% of the stance phase), and Fz dropped back to zero after reaching a second peak (Fzmax2 = 3.89±0.19N/kg) at 61±1% of the stance phase. The Iz was 1.42±0.13Ns/kg.

At the beginning of the stance phase of the hindlimbs in cats, a braking force was found, which reached a maximum (Fymax =0.81±0.21N/kg) at 17±0% of the stance phase, with a tangent β of 1.19±0.14 (Fig. 3). 36±1% of the stance phase was devoted to the braking phase, and thus 64% to the propulsive phase. Maximum propulsive force (Fymin = -1.08±0.13N/kg), was reached at 72±0% of the stance phase. The braking impulse (Iy+) was 0.07±0.02Ns/kg and the propulsive impulse (Iy-) was -0.18±0.03Ns/kg.

At the beginning of the stance phase, Fx is directed laterally (- for the right leg and + for the left leg; Fig. 2), (Fxmin = -0.02±0.06N/kg at 4±0% of the stance phase), then it moved to neutral (Fx =0) at 13% of the stance phase, maximum medial direction (Fxmax1 = 0.27±0.12N/kg at 22% of the stance phase), with a tangent γ of 1.19±0.14, and ended at 0.19±0.13N/kg, at the end of the braking phase (Fig. 3). During the propulsive phase, Fx+ of the same leg increased from 0.17±0.13N/kg (Fxdip) at 43±1% of the stance phase to 0.27±0.12N/kg at 69±0% of the stance phase, and then decreased back to a neutral position.

174 Ground reaction forces of walking cats; assessment and comparison with walking dogs

Video data Cats show a typical crouched position of the hindlimbs with marked flexion of the stifle and tibiotarsal joints (for videos see Appendix A). High speed video analysis revealed that cats supinate their front paws during the swing phase, and start to pronate just prior to the stance phase. Also diagonality in footfalls (putting the front paw in front of the contralateral front paw) was observed in these cats. Pronation and supination action were hardly present in the hindlimbs. For a more detailed description and discussion of the video data see Appendix A.

Cats vs. dogs Force plate data Ground reaction forces of cats were comparable to those of dogs (Compare Tables 1 and 2; Figs. 2 and 3). There were no significant differences in normalized peak forces for Fzmax1, Fymax, and also not for the time point at which forces cross zero (Fy=0) for the front and hindlimbs, respectively. Significant differences were found in Fzmax2 (p<0.001), Fymin (p<0.001), and Iy- (p<0.001) for the hindlimbs. During the propulsive phase, Fz in the hindlimbs of cats (Fzmax2 = 3.89±0.19N/kg) was significantly higher (p<0.001) compared to dogs (Fzmax2 = 3.03±0.16 N/kg). In cats, Fz dip (p<0.001) and Fzmax2 in the forelimbs (p=0.001 (left) and p=0.007 (right)), and Fymin (p<0.001) and Fzmax2 in the hindlimbs (p<0.001), occurred significantly earlier in the stride cycle, compared to dogs. Cats have similar Fzmax1 in the forelimbs compared to dogs, however the Iz in the forelimbs of cats 11 was significantly lower (p<0.001) due to a significantly shorter duration of the stance phase (Ts = 0.46±0.04s in cats, and 0.56±0.02s in dogs, respectively (p<0.001)). Cats had lower Fx+ forces, compared to dogs indicating less lateral directed action in cats. In contrast, negative Fx forces were observed, indicating medially directed action, whereas in dogs only positive Fx forces were found. However, Fx force differences were not highly significant, likely because there was a lot of variation in Fx forces. The tangent α and β in the forelimbs of cats was larger compared to dogs.

Video data Comparison of video images revealed that cats have more flexion of the stifle and tibiotarsal joints during gait compared to dogs, which is indicative of a stilted gait (for videos see Appendix A). Dogs walk in straight lines by putting the hind paw behind the ipsilateral front paw. Cats end their stance phase by initiating supination, while dogs start their supination during the swing phase. For a more detailed description and discussion of the video data see Appendix A.

Discussion

In this study, we described the stance phase of walking in the cat by using force plate analysis. Selection of 10 stride cycles based on speed consistency of gait resulted in very reliable data, as indicated by the high intraclass correlation coefficients. The variation in walking velocity of the cats in our study was small (0.6-0.8 m/s), and thus, velocity did not

175 Part II

cause a bias in our data. The force-time wave-forms demonstrated in this study can thus be used to evaluate lameness. Furthermore, force plate analysis may be used to monitor effects of treatment in cats with OA. A recent study demonstrated that measurement of Fzmax could be used to discriminate OA from non-OA cats, but failed to demonstrate treatment effects (Guillot et al. 2013). To assess ground reaction forces in cats requires an adaptation of force plates by the use of an overlay plate as described in this study. The effects of the overlay on force plate measurements were minimized by performing calibration of the force plate after attachment of the overlay. Limitations of force plate analysis are the need for expensive, specialized equipment and substantial expertise and effort from support personnel. In addition, force plate analysis evaluates an animal at one specific time point, outside of its normal environment, which may impact the generalization of the obtained results to the animal’s home environment (Brown et al. 2013). The use of force plate in client-owned cats can thus be challenging, but with some patience and training, most cats can be trained to walk on a leash. Another limitation of force plate analysis is that it only collects data during the stance phase. Several other techniques are described for collection of data during the swing phase like Pro-reflex® system (Rodriguez et al. 2013) and imaging of markers with (high speed) cameras (Lee et al. 2012). The force plate data demonstrate the resultant of different forces and do not provide insights in the individual vectors that contribute to this resultant.

11 Kinetic analyses with different techniques have been used to evaluate lameness and to monitor effects of treatment in several species. Force plate analysis has several advantages over other techniques. Force plate analysis generates a more comprehensive data set compared to pressure-sensitive walkway, in which only Fzmax is assessed (Lascelles et al. 2006, Guillot et al. 2012, Kano et al. 2013, Guillot et al. 2013). However, a pressure-sensitive walkway allows for measuring in less well trained animals, can be moved to different locations, and can be used to measure body weight distribution in standing position, which makes this easier to use in cats compared to force plate analysis (Guillot et al. 2013). Imaging with high speed cameras with reflective markers has several limitations: it is labor intensive, costly and not very accurate, especially in the proximal region of the extremities (shoulder, elbow, knee, and hip), because the markers are generally attached to the skin and thus move over the joint (Schwencke et al. 2012). Force plate analysis has the benefit of measuring Fy and Fx forces, together with Fz. Force-time wave-forms are also obtained from force plate analysis, although rarely used for kinetic analysis (Gregor et al. 2006, Maas et al. 2009), whereas analysis of complete wave-forms in all three directions is the established method for kinetic evaluation in human kinetic research (Fransz et al. 2013, Al-Nadaf et al. 2012). It is also possible to detect differences between complete gait wave-forms when no differentiation was detected by standard evaluation of single peak and impulse data (Al-Nadaf et al. 2012).

Force plate analysis has become the “gold standard” to evaluate effects of medical and surgical treatments of dogs with lameness objectively (Hazewinkel et al. 2003, Theyse et al. 2000) and may, thus, also be a useful method for cats. Fy has been studied to evaluate

176 Ground reaction forces of walking cats; assessment and comparison with walking dogs

effect of lumbosacral decompressive surgery in dogs with degenerative lumbosacral stenosis (Van Klaveren et al. 2005). Such an analysis might also be used in cats, as lumbar and lumbosacral segments are indicated as being the most painful areas of osteoarthritis- related pain in cats during orthopedic or radiological examination (Kranenburg et al. 2012, Lascelles et al. 2012). Force plate analysis in cats during downslope walking may aid in finding subtle differences in braking forces (Fymax), while upslope walking may aid in finding subtle differences in propulsive forces (Fymin) as well as vertical forces (Fz), as suggested by Gregor et al. (2006).

Fz, Fymax, and Fymin are reduced in cats after anterior cruciate ligament transaction (ACLT), which restored to pre-surgical values after an adaptation period, demonstrating the objective evaluation of subtle changes in cats by force plate analysis (Suter et al. 1998). This is in accordance with findings in Labradors with ACLT (Frost-Christensen et al. 2008). According to Moreau et al. (2013) the most consistent measurement of Fz can be obtained by using the lowest peak vertical force during stair climbing exercise, which has the potential to monitor subtle changes as a result of treatment of OA in cats.

Ground reaction forces during walking demonstrate a lot of similarities between cats and dogs. The forelimbs of cats steer the animal and they are more important than the hindlimbs for weight bearing and braking, which is demonstrated by a higher Fzmax1, Iz, Fymax2, and Iy+ in the forelimbs, compared to the hindlimbs and is similar in dogs 11 (Vilensky 1987). The braking phase of the forelimbs takes 57±1% of the stance phase (t Fy=0), and thus the propulsive phase takes 43±1% of the stance phase, indicating that the braking phase in the forelimbs is longer, whereas in the hindlimbs, the propulsive phase is longer. This is comparable to dogs.

The diagonality in footfalls that was observed on videos of walking cats and not of walking dogs, is in accordance with the findings of Bishop et al. (2008), and can be explained by the more crouched posture of cats during gait, compared to dogs. The propulsive impulse (Iy-) executed by the hindlimbs of cats is higher compared to that of dogs, especially during the beginning of the propulsive phase. The higher Iy- is caused by more flexion of the stifle and tibiotarsal joints in the hindlimbs in cats compared to dogs. This more crouched position during gait, thus a more compliant gait, requires more propulsive force (Bishop et al. 2008) compared to the stiffer gait of dogs, especially in the hind limbs.

Subtle differences in the gait between cats and dogs reveal a higher relative velocity compared to bone length, as well as the more flexed position of hindlimbs of cats compared to that of dogs. This is likely to account for the higher Fz forces in the hindlimbs of cats compared to dogs during the propulsive phase. In cats Fzdip and Fzmax2 in the forelimbs, as well as Fymin and Fzmax2 in the hindlimbs, occur significantly earlier in the stride cycle compared to dogs. This difference in timing is reflected by the start of supination during the end of the stance phase in cats, while in dogs, supination starts in

177 Part II

the swing phase. The supination is also responsible for different force-time wave-forms of Fx in cats compared to dogs. The tangent α and β in the forelimbs of cats is larger compared to dogs, which can be explained by the tiptoeing gait of cats compared to the more scuffing gait of dogs; especially in the forelimbs. Cats have similar Fzmax1 in the forelimbs compared to dogs; however the Iz in the forelimbs of cats is significantly lower due to a significantly shorter duration of the stance phase.

Conclusions

Force plate analysis is a useful tool for assessment of locomotion in cats. It can be applied to client-owned cats in the clinical setting and it is a valuable research tool for longitudinal follow up. It provides a non-invasive and objective assessment of biomechanical characteristics of gait with high reproducibility and provides more insight in kinetic characteristics compared to other methods of gait analysis. Cats differ from dogs in their force-time wave-forms in all three directions. This can be explained by the more crouched position of cats during gait, and by the difference in paw supination- pronation.

Conflict of interest statement

11 None of the authors has any financial or personal relationships that could inappropriately influence or bias the content of the paper.

Acknowledgements

The authors wish to thank Inge van Duiven and Harry van Engelen for taking care of the animals, and Lucas van der Wurff for guiding the animals over the force plate. Dr. Henri Heuven is acknowledged for his advises in the statistical analysis.

Appendix A: Supplementary material

Supplementary data associated with this article can be found, in the online version, at doi: .

178 Ground reaction forces of walking cats; assessment and comparison with walking dogs

References • Brown, D.C., Boston, R.C., Farrar, J,T., 2013. Comparison of force plate analysis and owner • Al-Nadaf, S., Torres, B.T., Budsberg, S.C., 2012. assessment of pain using the canine brief pain Comparison of two methods for analyzing inventory in dogs with osteoarthritis. Journal kinetic gait data in dogs. American Journal of of Veterinary Internal Medicine 27, 22-30. Veterinary Research 73, 189-193. • Clarke, S.P., Mellor, D., Clements, D.N., • Bénard, M.R., Jaspers, R.T., Huijing, P.A., Becher, Gemmill, T., Farrell, M., Carmicheal, S., Bennett, J.G., Harlaar, J., 2010. Reproducibility of D., 2005. Prevalence of radiographic signs hand held ankle dynamometry to measure of degenerative joint disease in a hospital altered ankle moment-angle characteristics population of cats. Veterinary Record 157, in children with spastic cerebral palsy. Clinical 793-799. Biomechanics 25, 802-808. • Corbee, R.J., Barnier, M.M.C., Lest C.H.A. van • Benito, J., DePuy, V., Hardie, E., Zamprogno, de, Hazewinkel H.A.W., 2013. The effect H., Thomson, A., Simpson, W., Roe, S., Hansen, of dietary long-chain omega-3 fatty acids B., Lascelles, B.D.X., 2013a. Reliability and supplementation on owner’s perception discriminatory testing of a client-based of behavior and locomotion in cats with metrology instrument, feline musculoskeletal naturally occurring osteoarthritis. Journal of pain index (FMPI) for the evaluation of Animal Physiology and Animal Nutrition 97, degenerative joint disease-associated pain in 846–853. cats. The Veterinary Journal 196, 368-373. • Dueland, R., Bartel, D.L., Antonson, E., 1977. • Benito, J., Gruen, M.E., Thomson, A., Simpson, Force plate technique for canine gait analysis: 11 W., Lascelles, B.D.X., 2012. Owner-assessed preliminary report on total hip and excision indices of quality of life in cats and the arthroplasty. Bulletin of the Hospital for Joint relationship to the presence of degenerative Diseases 38, 35-36. joint disease. Journal of Feline Medicine and • Fransz, D.P., Huurnink, A., Kingma, I., Verhagen, Surgery 14, 863-870. E.A., van Dieën, J.H., 2013. A systematic • Benito, J., Hansen, B., Depuy, V., Davidson, review and meta-analysis of dynamic tests G.S., Thomson, A., Simpson, W., Roe, S., and related force plate parameters used to Hardie, E., Lascelles, B.D.X., 2013b. Feline evaluate neuromusculoskeletal function musculoskeletal pain index: Responsiveness in foot and ankle pathology. Clinical and testing of criterion validity. Journal of Biomechanics 28, 591-601. Veterinary Internal Medicine 27, 474-482. • Frost-Christensen, L.N., Mastbergen, S.C., • Bennett, D., Morton, C., 2009. A study of Vianen, M.E., Hartog, A., DeGroot, J., Voorhout, owner observed behavioural and lifestyle G., van Wees, A.M.C., Lafeber, F.P.J.G., changes in cats with musculoskeletal disease Hazewinkel, H.A.W., 2008. Degeneration, before and after analgesic therapy. Journal of inflammation, regeneration, and pain/ Feline Medicine and Surgery 11, 997-1004. disability in dogs following destabilization or • Bishop, K.L., Pai, A.K., Schmitt, D., 2008. Whole articular cartilage grooving of the stifle joint. body mechanics of stealthy walking in cats. Osteoarthritis and Cartilage 16, 1327-1335. PlosOne 3, 1-6. • Gillette, R.L., Angle, T.C., 2008. Recent developments in canine locomotor analysis: a review. The Veterinary Journal 178, 165-176.

179 Part II

• Godfrey, D.R., 2005. Osteoarthritis in cats: A 1997). Journal of the American Veterinary retrospective radiological study. Journal of Medical Association 220, 628–632. Small Animal Practice 46, 425-429. • Hazewinkel, H.A.W., Brom, W.E. van den, • Grand, L., Ftomov, S., Timofeev, I., 2013. Theyse, L.F.H., Pollmeier, M., Hanson, P.D., Long-term synchronized electrophysiological 2003. Reduced dosage of ketoprofen for the and behavioral wireless monitoring of freely short-term and long-term treatment of joint moving animals. Journal of Neuroscience pain in dogs. Veterinary Record 152, 11-14. Methods 212, 237-241. • Kalis, R.H., Liska, W.D., Jankovits, D.A., 2012. • Gregor, R.J., Smith, D.W., Prilutsky, B.I., 2006. Total hip replacement as a treatment option Mechanics of slope walking in the cat: for capital physeal fractures in dogs and cats. Quantification of muscle load, length change, Veterinary Surgery 41, 148-155. and ankle extensor EMG patterns. Journal of • Kano, W., Rahal, S., Mesquita, L., Agosthino, Neurophysiology 95, 1397-1409. F., Faria, L., 2013. Gait analysis in a cat with • Grösslinger,K., Lorinson, D., Bockstahler, B., scapula luxation and contralateral forelimb Malleczek, D., Tichy, A., Windischbauer, G., amputation. The Canadian Journal 54, 990- 2006. Postoperative examination of scapular 991. neck fractures in 4 cats with goniometry • Kranenburg, H.C., Meij, B.P., van Hofwegen, and force plate analysis. Wiener Tierarztliche E.M.L., Voorhout, G., Slingerland, L.I., Picavet, Monatsschrift 93, 170-175. P., Hazewinkel, H.A.W., 2012. Prevalence of • Guedes, A.G.P., Matthews, N.S., Hood, D.M., spondylosis deformans in the feline spine and 11 2012. Effect of ketamine hydrochloride on the correlation with owner-perceived behavioural analgesic effects of tramadol hydrochloride changes. Veterinary and Comparative in horses with signs of chronic laminitis Orthopaedics and Traumatology 25, 217-223. associated pain. American Journal of • Lascelles, B.D.X, Roe, S.C., Smith, E., Reynolds, Veterinary Research 73, 610-619. L., Markham, J., Marcellin-Little, D., Bergh, • Guillot, M., Moreau, M., Anjou, M.A. d’, Martel- M.S., Budsberg, S.C., 2006. Evaluation of a Pelletier J., Pelletier J.P., Troncy, E., 2012. pressure walkway system for measurement of Evaluation of osteoarthritis in cats : novel vertical limb forces in clinically normal dogs. information from a pilot study. Veterinary American Journal of Veterinary Research 67, Surgery 41, 328-335. 277-282. • Guillot, M., Moreau, M., Heit, M., Martel- • Lascelles, B.D.X., Findley, K., Correa, M., Pelletier, J., Toncy, E., 2013. Characterization Marcellin-Little, D., Roe, S., 2007. Kinetic of osteoarthritis in cats and meloxicam evaluation of normal walking and jumping efficiency using objective pain evaluation in cats, using a pressure-sensitive walkway. tools. The Veterinary Journal 196, 360-367. Veterinary Record 160, 512-516. • Halbertsma, J.M., 1983. The stride cycle • Lascelles, B.D.X., 2010. Feline degenerative of the cat: the modeling of locomotion joint disease. Veterinary Surgery 39, 2-13. by computerized analysis of automatic • Lascelles, B.D.X., DePuy, V., Thomson, A., recordings. Acta Physiologica Scandinavia Hansen, B., Marcellin-Little, D.J., Biourge, V., Supplementum 521, 1-75. Bauer, J.E., 2010. Evaluation of a therapeutic • Hardie, E.M., Roe, S.C., Martin, F.R., 2002. diet for Feline degenerative joint disease. Radiographic evidence of degenerative joint Journal of Veterinary Internal Medicine 24, disease in geriatric cats: 100 cases (1994– 487-495.

180 Ground reaction forces of walking cats; assessment and comparison with walking dogs

• Lascelles, B.D.X., Dong, YH., Marcellin-Little, • Oomen, A.M., Oosterlinck, M., Pille, F., D.J., Thomson, A., Wheeler, S., Correa, M., 2012. Sonneveld, D.C., Gasthuys, F., Back, W., 2012. Relationship of orthopedic examination, Use of a pressure plate to analyse the toe-heel goniometric measurements, and radiographic load redistribution underneath a normal signs of degenerative joint disease in cats. shoe and a shoe with a wide toe in sound BMC Veterinary Research 8, 10. warmblood horses at the walk and trot. • Lee, H.-Y., Hsieh, T.-H., Liang, J.-I., Yeh, M.-L., Research in Veterinary Science 93, 1026-1031. Chen, J.-J.J., 2012. Quantitative video-based • Prilutsky, B.I., Maas, H., Bulgakova, M., Hodson- gait pattern analysis for hemiparkinsonian Tole, E.F., Gregor, R.J., 2011. Short-term motor rats. Medical and Biological Engineering and compensations to denervation of feline Computing 50, 937-946. soleus and lateral gastrocnemius result in • Maas, H., Gregor R.J., Hodson-Tole, E.F., Farrell, preservation of ankle mechanical output B.J., Prilutsky, B.I., 2009. Distinct muscle during locomotion. Cells Tissues Organs 193, fascicle length changes in feline medial 310–324. gastrocnemius and soleus during slope • Rialland, P., Bichot, S., Moreau, M., Guillot, walking. Journal of Applied Physiology 106, M., Lussier, B., Gauvin, D., Martel-Pelletier J., 1169-1180. Pelletier, J.P., Toncy, E., 2012. Clinical validity of • Mazurek, K.A., Holinski, B.J., Everaert, D.G., outcome pain measures in naturally occurring Stein, R.B., Etienne-Cummings, R., Mushahwar, canine osteoarthritis. BMC Veterinary V.K., 2012. Feed forward and feedback control Research 8, 162. 11 for over-ground locomotion in anaesthetized • Rodriguez, E.B., Chagas, P.S.C., Silva, P.L.P., cats. Journal of Neural Engineering 9, Article Kirkwood, R.N., Mancini, M.C., 2013. Impact of number 026003. leg length and body mass on the stride length • McLaughlin Jr., R.M., Miller, C.W., Taves, C.L., and gait speed of infants with normal motor Hearn, T.C., Palmer, N.C., Anderson, G.I., 1991. development: A longitudinal study. Brazilian Force plate analysis of triple pelvic osteotomy Journal of Physical Therapy 17, 163-169. for the treatment of canine hip dysplasia. • Schwencke, M., Smolders, L.A., Bergknut, N., Veterinary surgery 20, 291-297. Gustås, P., Meij, B.P., Hazewinkel, H.A.W., 2012. • Merkens, H.W., 1987. Quantitative evaluation Soft tissue artifact in canine kinematic gait of equine locomotion using force plate data. analysis. Veterinary Surgery 41, 829-837. Thesis Utrecht University 11-12. • Slingerland, L.I., Hazewinkel, H.A.W., Meij, B.P., • Moreau, M., Guillot, M., Pelletier, J.P., Martel- Picavet, P., Voorhout, G., 2011. Cross-sectional Pelletier, J., Troncy, E., 2013. Kinetic peak study of the prevalence and clinical features vertical force measurement in cats afflicted of osteoarthritis in 100 cats. The Veterinary by coxarthritis : data management and Journal 187, 305-309. acquisition protocols. Research in Veterinary • Smolders, L.A., Voorhout, G., Ven, R. van de, Science 95, 219-224. Bergknut, N., Grinwis, G.C.M., Hazewinkel, • Mueller, M., Bockstahler, B., Skalicky, M., H.A.W., Meij, B.P., 2012. Pedicle screw-rod Mlacnik, E., Lorinson, D., 2007. Effects of radial fixation of the canine lumbosacral junction. shockwave therapy on the limb function Veterinary Surgery 41, 720-732. of dogs with hip osteoarthritis. Veterinary Record 160, 762-765.

181 Part II

• Spaak, B., van Heel, M.C.V., Back, W., 2013. Toe • Zamprogno, H., Hansen, B.D., Bondell, modifications in hind feet shoes optimise H.D., Sumrell, A.T., Simpson, W., Robertson, hoof unrollment in sound Warmblood horses I.D., Brown, J., Pease, A.P., Roe, S.C., Hardie, at trot. Equine Veterinary Journal 45, 485-489. E.M., Wheeler, S.J., Lascelles B.D.X., 2010. • Suter, E., Herzog W., Leonard, T.R., Nguyen, Item generation and design testing of a H., 1998. One-year changes in hind limb questionnaire to assess degenerative joint kinematics, ground reaction forces and disease-associated pain in cats. American knee stability in an experimental model of Journal of Veterinary Research 71, 1417-1424. osteoarthritis. Journal of Biomechanics 31, 511-517. • Theyse, L.F.H., Hazewinkel, H.A.W., van den Brom, W.E., 2000. Force plate analyses before and after surgical treatment of unilateral fragmented coronoid process. Veterinary and Comparative Orthopaedics and Traumatology 13, 135-140. • Van der Peijl, G.J.W., Schaeffer, I.G.F., Theyse, L.F.H., Dijkshoorn, N.A., Schwencke, M., Hazewinkel, H.A.W., 2012. Osteochondrosis dissecans of the tarsus in Labrador retrievers: 11 Clinical signs, radiological data and force plate gait evaluation after surgical treatment. Veterinary and Comparative Orthopaedics and Traumatology 25, 126-134. • Van Klaveren, N.J., Suwankong, N., Boer, S. de, Brom, W.E. van den, Voorhout, G., Hazewinkel, H.A.W., Meij, B.P., 2005. Force plate analysis before and after dorsal decompression for treatment of degenerative lumbosacral stenosis in dogs. Veterinary Surgery 34, 450-456. • Vilensky, J.A., 1987. Locomotor behavior and control in human and non-human primates: comparisons with cats and dogs. Neuroscience & Biobehavioral Reviews 11 263-274.

182 Ground reaction forces of walking cats; assessment and comparison with walking dogs

Appendix A Video data Because of intersubject similarity, only the video of one cat and one dog are shown.

Video data of cats High speed video analysis revealed that cats supinate their paws during the swing phase, and start to pronate just prior to the stance phase. Cats still pronate their front paws during the beginning of the stance phase, until the paw reaches a neutral position (Video 1 and 2). The most lateral phalanx touches the ground first, and then the other phalanges touch the ground. At the end of the stance phase the paw is supinated gradually with the medial phalanx released first, followed by the other phalanges. Also diagonality in footfalls (putting the front paw in front of the contralateral front paw) was observed in these cats.

All cats show a typical crouched position of the hindlimbs with marked flexion of the stifle and tibiotarsal joints (Video 1). When a hindlimb starts the braking phase, the paw is still moving medially. The hindlimb then reaches a neutral (Fx = 0) position, until it starts the final part of the propulsive phase in which the hindlimb has to be lifted and moved laterally, with the medial phalanx released first, followed by the other phalanges. Diagonality in footfalls was observed also (Video 3). Pronation and supination movements were hardly observed in the hindlimbs. 11 Video data of cats compared to dogs Comparison of video images revealed that cats have more flexion of the stifle and tibiotarsal joints during gait compared to dogs, which is indicative of a stilted gait (Video 1 and 4). Dogs walk in straight lines by putting the hind paw behind the ipsilateral front paw (Video 5). Cats supinate their paws during the swing phase, and turn them to neutral position during the beginning of the stance phase (Video 1). Cats first put their most lateral phalanx on the ground, followed by the more medial phalanges, whereas dogs put all four weight bearing phalanges on the ground at the same time. Cats end their stance phase by initiating supination, while dogs start their supination during the swing phase. At the beginning of the swing phase, the dog supinates its legs and then turns it back in neutral position during the swing phase. Dogs end the stance phase by moving the paw medially, to be able to start the supination in the swing phase. On Video 6, it can be seen that dogs have almost no supination action in the hindlimbs. The above described differences in kinematic patterns of the fore- and hindlimbs may explain (part of) the differences in ground reaction forces observed in the present study.

183 184 Chapter 12

Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

Ronald J. Corbee Marianna A. Tryfonidou Guy C.M. Grinwis Baukje A. Schotanus Martijn R. Molenaar George Voorhout Arie B. Vaandrager Herman A.W. Hazewinkel

Submitted to The Veterinary Journal

185 Part II

Abstract

The first aim of the study was to determine whether vitamin D supplementation influences the effects of high vitamin A intake on new bone formation in adult cats. The second aim was to determine whether high vitamin A intake in cats can induce liver pathology and if so, whether the current safe upper limit for dietary intake of vitamin A for healthy adult cats is adequate for prevention of skeletal and liver pathology. To this end, 24 healthy adult cats were divided into four groups receiving a control diet supplemented with peanut oil (CTR), or peanut oil containing a 100-fold increased amount of vitamin A (HA), or a 100-fold increased amount of vitamin A and a 5-fold increased amount of vitamin D (HAMD), or a 100-fold increased amount of vitamin A and a 65-fold increased amount of vitamin D (HAHD) during 18 months. Cats consuming vitamin A at the current upper limit (NRC 2006) neither showed abnormal locomotion nor clinical symptoms of liver failure after 18 months of supplementation. However, the subtle skeletal changes and the liver pathology in these cats indicate that current vitamin A NRC recommendations for cats are too high. The addition of vitamin D does not seem to influence bone pathology in cats during the 18 month study period. Moderate elevated dietary 12 vitamin D levels (HAMD) may be protective for development of liver pathology in cats consuming large amounts of vitamin A, but higher dietary levels of vitamin D (HAHD) are no longer protective.

186 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

Introduction

Hypervitaminosis A in cats has been reported since 1957 (Christi, 1957) and is mostly associated with chronic consumption of large amounts of raw liver. Most common pathological findings of hypervitaminosis A in cats are skeletal hyperostosis of the vertebrae and periarticular hyperostosis (Polizopoulou et al. 2005). In kittens, teratogenic effects of high vitamin A intake include cleft palate, craniofacial and pelvic abnormalities, shallow ophthalmic orbits, and vascular abnormalities (Freytag et al. 2003). These effects were demonstrated when queens were fed a diet with 1-2 times the safe upper limit according to the NRC requirements, for 2-3 years prior to and during pregnancy (NRC 2006). In other species, such as rat, mice, pigs, calves, and dogs, microphthalmia and developmental problems of the eye as well as premature closure of growth plates are reported in case of hypervitaminosis A (Cho et al. 1975, Nakane et al. 1978, Reiner et al. 2004, Pasquinelli et al. 2007, Rothenberg et al. 2007).

In the skeleton, vitamin A promotes the differentiation of mesenchymal stem cells into osteoblasts and osteoclasts, as well as cartilage maturation and mineralization (Jimenez et al. 2001). Cartilage maturation and mineralization are also stimulated by vitamin D (DeLuca et al. 1982). In adult humans and adult rats, high vitamin A intake is associated with osteoporosis, indicating a shift towards more osteoclasia (Lind et al. 2012, Hu et al. 2010). Only a few case reports in men demonstrate cervical hyperostosis due to high vitamin A intake (Wendling et al. 2009). In adult cats however, hypervitaminosis A is characterized by new bone formation, indicating a shift towards more osteoblast 12 activity and bone formation (Franch et al. 2000). However, excessive vitamin A intake alone did not cause the typical skeletal new bone formation in cats after 2 years (Freytag et al. 2003). The consumption of raw liver, containing both excessive vitamin A and more vitamin D compared to most other foods (Becker and Kienzle 2013) or supplements containing vitamin A and D were suggested to be the cause of this new bone formation (Polizopoulou et al. 2005, Goedegebuure and Hazewinkel 1981). Therefore the first aim of the study was to determine whether vitamin D supplementation influences the effects of vitamin A supplementation on the skeleton of adult cats. We hypothesize that skeletal hyperostosis found in hypervitaminosis A in cats is the result of excessive intake of vitamin A in combination with vitamin D supplementation.

Generally hypervitaminosis A in humans is characterized by liver fibrosis and pupil edema (Nollevaux et al. 2006, Stickel et al. 2011); pathologies not reported in cats (Freytag et al. 2003). However, fatty infiltration of Kupffer cells and giant cells were demonstrated in cats consuming large amount of vitamin A (Seawright et al. 1967). Vitamin A is mostly stored in the hepatic stellate cells in both humans (Shirakami et al. 2012) and cats (Seawright et al. 1967). High vitamin A intake activates these hepatic stellate cells and finally results in the loss of hepatic stellate cell vitamin A stores due to over-accumulation and leakage of vitamin A into the parenchyma. Both activation of hepatic stellate cells and leakage of vitamin A into the liver parenchyma result in changes in extracellular matrix deposition

187 Part II

(mainly collagen I) leading to the onset of liver fibrosis in humans (Shirakami et al. 2012). Therefore, the second aim of this study was to determine whether high vitamin A intake in cats can induce liver pathology and if so, whether the current upper limit (NRC 2006) for dietary intake of vitamin A for healthy adult cats is safe for prevention of skeletal and liver pathology.

Materials and methods Ethical statement All experiments were approved by the Animal Welfare Committee on Experimental Animal Use, as required by Dutch legislation (DEC 2011.III.01.008).

Animals, foods, and supplements Twenty-four healthy, neutered cats (12 males, 12 females, age 5.5 (range 1.5-12.0) years, body weight (BW) 4.8 (3.0-6.9) kg, body condition score of 6±1 on a 9-point scale were randomly divided into four groups. According to Dutch Animal Welfare legislation, the cats were group-housed with 6 cats per group. All cats were fed a commercial diet for the prevention of struvite urolithiasis (Hill’s Prescription DietTM c/dTM Multicare Feline, chicken, dry). The amount fed per cat was calculated by 1.4x70x(BW in kg)0.75 in kcal Metabolizable Energy (ME) per day. Body weight of the individual cats was recorded every two weeks, and the amount of food was adjusted if necessary to maintain the cats on their normal body weight and body condition score. Diet composition and daily intake of the cats of the different groups are demonstrated in Table 1 and were close to the current upper 12 limit (NRC 2006) for vitamin A and vitamin D in all supplemented groups. All cats received daily an addition of 0.5 mL peanut oil (8.84 kcal per mL) mixed with the food. The peanut oil contained 53,393 IU of vitamin A per mL in the high vitamin A group (HA); 53,393 IU of vitamin A and 250 IU of vitamin D per mL in the high vitamin A moderate vitamin D group (HAMD); and 53,393 IU of vitamin A and 3,983 IU of vitamin D per mL in the high vitamin A and high vitamin D group (HAHD). The control group received 0.5 mL plain peanut oil.

Blood measurements At the start of the study, blood samples from the 24 cats were drawn to determine general health status of the animals, including complete blood count, blood urea nitrogen, creatinine, alkaline phosphatase, bile acids, glucose, total protein, albumin, electrolytes, calcium, phosphate, and thyroid hormone. Blood coagulation was determined by evaluating the intrinsic pathway (activated partial thromboplastin time [APTT]), the extrinsic pathway (prothrombin time (PT), fibrinogen content), and platelet count. Cats were excluded if abnormalities were found. In addition, plasma 25-hydroxyvitamin D (25OHD), retinol, and retinyl esters were determined. Every 6 months, until the end of the study at 18 months, blood samples were drawn for determination of calcium, phosphate, total protein, albumin (using SynchronCX® systems on the Beckman Coulter; Brea, CA, USA), 25OHD, retinol, and retinyl esters (after hexane-extraction by High-performance liquid chromatography and mass spectrometry (HPLC-MS) as described below).

188 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

Table 1: Correlation coefficients and symmetry index of force plate analysis in cats

Per 100 g Per 100 g dry Per 100 as fed matter kcal ME Crude protein (g) 32.6 34.5 8.4 Crude fat (g) 15.6 16.5 4 NFE (g) 40.6 43 10.5 Crude fiber (g) 0.7 0.8 0.2 Moisture (g) 5.5 Calcium (g) 0.71 0.75 0.18 Phosphorus (g) 0.65 0.69 0.17 Vitamin A (IU) 313 332 81 Vitamin D (IU) 39 41 10 Vitamin E (mg) 55 58 14

Metabolizable Energy kcal / 100g 389 412 kJ / 100g 1627 1722

Supplement high vitamin A group (HA) contained 53,393 IU retinol per mL Supplement high vitamin A, moderate vitamin D group (HAMD) contained 53,393 IU retinol and 250 IU cholecalciferol per mL Supplement high vitamin A, high vitamin D group (HAHD) contained 53,393 IU retinol and 3,983 IU cholecalciferol per mL 12

Daily intake average cat Control HA HAMD HAHD Safe upper limit according to NRC 2006 Vitamin A (IU) 257 26,953 26,953 26,953 26,500 Vitamin D (IU) 31 31 156 2,022 2,385

NFE = nitrogen free extract

Table 2: Retinol and retinyl esters content of the liver in cats expressed in pmol per nmol cholesterol

0 18 CTR Retinol 14.9±23.9A 20.8±8.0A HA 86.8±77.0A 441.9±331.8B HAMD 10.7±3.6A 212.3±156.9B HAHD 18.6±35.0A 293.1±109.7B CTR = Control group; HA = High vitamin A group; CTR Retinyl esters 755±618A 808±633A HAMD = High vitamin A, moderate vitamin D group; HA 1393±756A 21193±6801B HAHD = High vitamin A, high vitamin D group; 0 = A HAMD 1680±1101A 17627±7939B Baseline; 18 = After 18 months of supplementation; HAHD 1650±3637A 18258±5506B = significantly different from B (p<0.05)

189 Part II

All HPLC-MS solvents were obtained from Biosolve (Valkenswaard, the Netherlands) with exception of chloroform (Carl Roth, Karlsruhe, Germany) and n-hexane (LabScan, Gliwice, Poland) and were of HPLC grade. Retinol, retinylacetate, and retinylpalmitate standards were obtained from Sigma-Aldrich (St. Louis, USA) and the internal standard 26,27-hexadeuterium-25-hydroxyvitamin D3 (D6-25OHD) was from Synthetica (Oslo, Norway).

Samples were kept under red light and all procedures were carried out in brown vials to prevent retinoid isomerization and oxidation. Hydrophobic vitamins were extracted from 0.2 mL plasma by subsequent additions of 0.25 mL demineralized water, 0.5 mL ethanol, containing 250 pmol retinylacetate and 150 pmol of D6-25OHD as internal standards, and 4 mL hexane containing 0.002% butylated hydroxytoluene (BHT) as an anti-oxidant. After each addition, samples were vortexed 3 times for 5 secs. After centrifugation (5 min, 1000g) the upper phase was transferred to a new brown glass tube and hexane was evaporated under nitrogen gas and stored at -20° C. HPLC separation was performed as described by Testerink et al. (2012). In short, the lipid fraction was reconstituted in 0.1 mL methanol/chloroform (1/1, v/v) and 10 µL was injected on a Lichrospher RP18-e column (5 μm, 250 x 4.6 mm; Merck, Darmstadt, Germany). A gradient was generated from acetonitrile:water 95/5 to acetone/chloroform 85/15, v/v, at a constant flow rate of 1 mL/min. Total run time per sample was 13 min. Mass spectrometry of lipids was performed using Atmospheric Pressure Chemical Ionization on a Biosystems API-4000 Q-trap (MDS Sciex, Concord, Canada). The system was controlled by Analyst version 12 1.4.2 software (MDS Sciex, Concord, ON, Canada) and operated in positive ion mode and in the multiple reaction monitoring (MRM) mode using the following settings: source temperature 420° C, nebulizer gas (GS1) 5, nebulizer current 3 μA, curtain gas 10, collision gas High and declustering potential and collision energy were empirically optimized for each compound. The MRM transitions (m/z) used were: 269.2 → 93.1 for retinoids (retinol and the different retinyl esters); 383.3 → 365.3 for 25OHD, 389.3 → 371.3 for D6-25OHD, and 369.3 → 161.1 for cholesterol.

Data analysis was performed using Analyst 1.4.2 software (MDS Sciex, Concord, ON, Canada) and quantitation of 25OHD, retinol and retinyl esters was done relative to their respective internal standards. The exact concentration of the internal standards was determined before each experiment by spectrophotometry (D6-25OHD at 264 nm using a molar absorption coefficient of 19,400 1M- cm-1 and retinyl acetate at 325 nm using a molar absorption coefficient of 52,480 1M- cm-1). The response factors of retinol and retinyl esters, in comparison to the internal standard retinyl acetate was determined by calibration curves run under similar conditions as the samples. The recovery was approximately 80% for both internal standards, and the intra assay variation was 10% for retinol and 8.5% for 25OHD at a respective plasma concentration of 750 nmol/L and 125 nmol/L.

190 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

Table 3: Radiography findings in cats

0 6 12 18 C1 Hip OA (1) Spondylosis Hip OA (1) Spondylosis Hip OA (1) Spondylosis Hip OA (1) Spondylosis L7-S1 L7-S1 L7-S1 L7-S1 C2 No abnormalities No abnormalities No abnormalities No abnormalities C3 Patellar luxation Elbow Patellar luxation Elbow Patellar luxation Elbow Patellar luxation Elbow OA (1) OA (1) OA (1) OA (1) C4 Tarsal OA (1) Tarsal OA (1) Tarsal OA (1) Tarsal OA (1) C5 Shoulder OA (1) Shoulder OA (1) Shoulder OA (1) Shoulder OA (1) C6 No abnormalities No abnormalities No abnormalities No abnormalities HA1 Spondylosis T11-T12- Spondylosis T2-T3- Spondylosis C2-C3, Spondylosis C2-C3, T13-L1-L2 T4-T5, T7, T11-T12- T1-T2-T3-T4-T5, T7, T1-T2-T3-T4-T5-T6-T7, T13-L1-L2 T11-T12-T13-L1-L2, T11-T12-T13-L1-L2, elbow OA (1) elbow OA (1) HA2 Spondylosis T11-T12 Spondylosis T11-T12 Spondylosis C2-3, T3- Spondylosis C2-3, T3- T4, T6-T7, T11-T12 T4, T6-T7-T8, T11-T12 HA3 No abnormalities No abnormalities No abnormalities No abnormalities HA4 No abnormalities No abnormalities No abnormalities No abnormalities HA5 No abnormalities No abnormalities Hip OA (1) Hip OA (1) HAMD1 No abnormalities No abnormalities No abnormalities Spondylosis T13-L1 HAMD2 Spondylosis C2-C3, Spondylosis C2-C3, T6- Spondylosis C2-C3, Spondylosis C2-C3, T11-T12 T7-T8-T9-T10-T11-T12 T3-T4-T5-T6-T7-T8-T9- T3-T4-T5-T6-T7-T8-T9- T10-T11-T12, Tarsal T10-T11-T12, Tarsal OA 12 OA (1) (1), Knee OA (2) HAMD3 No abnormalities No abnormalities Spondylosis T11- Spondylosis T11- T12-T13 T12-T13, Knee OA (1) HAMD4 No abnormalities No abnormalities No abnormalities No abnormalities HAMD5 No abnormalities No abnormalities No abnormalities No abnormalities HAHD1 Spondylosis T11-T12, Spondylosis T11- Spondylosis T11- Spondylosis T9-T10- Carpal OA (1), Knee T12-T13, Carpal OA (1), T12-T13, L7-S1, Carpal T11-T12-T13, L7-S1, OA (1) Knee OA (1) OA (1), Knee OA (1), Carpal OA (1), Knee OA Tarsal OA (1) (2), Tarsal OA (1) HAHD2 Spondylosis T6 Spondylosis T6, Tarsal Spondylosis T6, Tarsal Spondylosis T6, Tarsal OA (1) OA (1) OA (1) HAHD3 Spondylosis C2-C3, Spondylosis C2-C3, Spondylosis C2-C3, Spondylosis C2-C3, Knee OA (1) Knee OA (1) Knee OA (1) L6-L7, Knee OA (1) HAHD4 No abnormalities Knee OA (1) Knee OA (1) Knee OA (1), Hip OA (1) HAHD5 No abnormalities No abnormalities No abnormalities No abnormalities HAHD6 No abnormalities No abnormalities Tarsal OA (1) Tarsal OA (1)

C = Control group; HA = High vitamin A group; HAMD = High vitamin A, moderate vitamin D group; HAHD = High vitamin A, high vitamin D group; 0 = Baseline; 6 = After 6 months of supplementation; 12 = After 12 months of supplementation; 18 = After 18 months of supplementation; OA = Osteoarthritis; number between brackets indicates OA severity (i.e. 1 = mild, 2 = moderate, 3 = severe)

191 Part II

Fig. 1: Plasma retinol and 25-hydroxyvitamin D levels (expressed as mean±SEM) in cats during 18 months on a control diet supplemented with peanut oil (CTR, n=6), or peanut oil containing a 100-fold increased amount of vitamin A (HA, n=5), or a 100-fold increased amount of vitamin A and a 5-fold increased amount of vitamin D (HAMD, n=5), or a 100-fold increased amount of vitamin A and a 65-fold increased amount of vitamin D (HAHD, n=5). * = significantly different compared to the other groups (p<0.01)

192 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

Fig. 2: Histology and immunohistochemical stainings of the liver biopsies of 21 cats after 18 months on a control diet supplemented with peanut oil (CTR, n=6), or peanut oil containing a 100-fold increased amount of vitamin A (HA, n=5), or a 100-fold increased amount of vitamin A and a 5-fold increased amount of vitamin D (HAMD, n=5), or a 100-fold increased amount of vitamin A and a 65-fold increased amount of vitamin D (HAHD, n=5). 12 K19 = Cytokeratin 19 (a marker for progenitor cells); α-SMA = α-Smooth muscle actin (a general mesenchymal marker, staining activated stellate cells); SR = Picrosirius red (a collagen staining). Typical examples from each group are demonstrated for K19, α-SMA and SR

Table 4: Retinol and retinyl esters content of the liver in cats expressed in pmol per nmol cholesterol

K19 α-SMA por α-SMA par Ki67 hep Ki67 HSC SR par Hypertrophic HSC C 1.42±0.58A,B 1.33±0.52A 1.00±0.63A 3.50±3.67 0.67±0.82 1.17±0.75A 0.94±0.90A HA 1.90±0.42B 2.50±0.35B 2.40±0.89B 19.0±29.9 1.00±1.73 1.60±1.14B 20.0±9.22B HAMD 1.20±0.45A 1.60±0.55A,B 1.60±0.55A,B 3.60±2.30 0.00±0.00 2.00±1.00B 2.00±5.28B HAHD 1.92±0.49B 1.75±0.61B 2.00±0.84B 7.00±9.38 1.33±2.42 2.00±0.89B 22.2±8.92B

C = Control group; HA = High vitamin A group; HAMD = High vitamin A, moderate vitamin D group; HAHD = High vitamin A, high vitamin D group; 0 = Baseline; 18 = After 18 months of supplementation. K19 = Cytokeratin 19; α-SMA = α-Smooth muscle actin; por = Portal area; par = Parenchymal area; hep = Hepatocytes; HSC = Hepatic stellate cell; SR = Picrosirius red. Total number of K19 positive cells per microscopic field; percentage of α-SMA expression, total number of Ki67 positive cells per microscopic field; percentage of picrosirius red staining; total number of hypertrophic hepatic stellate cells per high power field; A = significantly different from B (p<0.05)

193 Part II

Radiographical examination At the start of the study, plain radiographs were taken of all the joints, long bones, and of the vertebral column with a digital radiography system (Philips digital Rad TH, Philips, the Netherlands). This was repeated every 6 months until the end of the study at 18 months. The radiographs were analyzed by a single radiologist (GV) who was blinded to the group divisions. Skeletal abnormalities including osteophytosis, new bone formation, and loss of mineralization of bone, were registered.

Force plate analysis At the start of the study, and every three months thereafter, all cats were evaluated by force plate analysis as described previously (Corbee et al. 2014). In brief: the body mass of each cat was determined on an electronic scale (DIWAC VS150) immediately before force plate measurement. A quartz piezoelectric force plate (Kistler type 9261) with Kistler 9865B charge amplifiers, mounted flush in a walk way of 5m was used. The force platform area, standard 40cm long and 60cm wide for dogs, was decreased to 25cm long and 60cm wide by a firmly attached overlay plate allowing for measurement of the cats. Sampling rate was 100Hz. Data per leg were collected in 10 steps in one session. All ground reaction force data were normalized to body mass. Impulses were calculated by NI Lab view 8.2 software.

Liver biopsies At the start and at the end of the study, liver biopsies were taken with the aid of a Pro. 12 MagTM 16 ga x 20 cm biopsy needle (Angiotech, Gainesville, FL, USA) with the technique as described previously (Rothuizen and Twedt 2009), under general anesthesia. Prior to obtaining liver biopsies, blood coagulation was determined as described above. The cats were pre-medicated with methadon (0.1 mg/kg BW), midazolam (0.2 mg/kg BW) and ketamine (5 mg/kg BW) together intramuscularly, induced with propofol (4 mg/kg BW) intravenously and maintained under anesthesia with propofol (5 mg/kg BW/h) by continuous rate infusion.

The liver biopsies were cut in half. One half of the biopsy was stained with haematoxylin and eosin (HE) and for immunohistochemically stained sections Cytokeratin 19 (K19: a marker for progenitor cells), Ki67 (a proliferation marker), α-Smooth muscle actin (α-SMA: a general mesenchymal marker, staining activated stellate cells), and Picrosirius red (SR: a collagen staining). All HE slides were blindly evaluated by a single observer (GG). The immunohistochemically stained sections were evaluated by the use of ImageJ software and the color deconvolution plugin (v1.47b Colour Convolution, NIH). For stromal and cytoplasmic stainings, the ratio of the stained area to the total area was determined and used for statistical analysis. For the nuclear Ki67 stainings, the number of stained cells was counted per microscopic field, as well as the number of swollen stellate cells visible per high power field on the HE slides.

194 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

The other half of the biopsy was used for quantification of retinol, retinyl esters, and 25OHD by mass spectrometry as described above and corrected for cholesterol content.

Statistical analysis Data obtained from evaluation of radiographs are presented as descriptive data. Data are presented as mean±sd for normal distributed data (mean±SEM in figure 1), or as median (range) for data that were not normally distributed.

A Kolomogorov-Smirnov test was used to test for normal distribution of the data. Differences between groups were tested with t-test for normally distributed data, and with a Mann-Whitney U-test for non-parametric data. Differences between time points were tested with multivariate analysis with the use of R-statistics software (R i386 3.0.1). The best model fit was used. p<0.05 was set as the level of significance for univariate analysis. p<0.01 was set as the level of significance for multivariate analysis to correct for multiple testing. p<0.10 was set as the level for tendency for univariate analysis.

Results

From the 24 cats that started the study, 2 did not reach the study endpoint. After 12 months one cat (female neutered, 12 years old, BW 4.8 kg, BCS 7 on a 9-point scale) in the HAMD group developed diabetes mellitus, and was excluded from the study. After 13 months one cat (female neutered, 12 years old, BW 3.0 kg, BCS 5 on a 9-point scale) in the HA group was diagnosed with a primary lung tumor, and was excluded from the study. 12 With oil supplementation, food intake was maintained. All 22 cats maintained their BW and BCS throughout the study.

Blood measurements did not differ between all groups. There were no significant differences in time, for plasma calcium, phosphate, total protein, and albumin for all groups. The means were for total calcium (2.60±0.14 mmol/L), phosphate (1.54±0.26 mmol/L), total protein (58±4 g/L), and albumin (27±2 g/L). Blood analysis revealed that the high vitamin A intake (i.e. in the HA, HAMD, and HAHD group) caused a gradual increase in plasma levels of retinol, resulting in a two-fold higher concentration after 18 months, whereas there was no significant increase in time in plasma levels of retinol in the control group (Figure 1, Appendix A). Plasma levels of retinyl esters varied considerably between the cats, and no significant differences were noted between groups (Appendix A). In the liver, both retinol and retinyl esters increased more than ten-fold in the HA, HAMD, and HAHD group compared to baseline and compared to the control group (Table 2). The increase of plasma levels of retinoids by the high dietary vitamin A intake (i.e. in the HA, HAMD, and HAHD group) were not affected by the additional intake of vitamin D (Figure 1, Appendix A). Vitamin D supplementation resulted in a significant increase in plasma 25OHD levels, approximately two-fold in the plasma of cats from the HAMD group and six-fold in the HAHD cats (Figure 1, Appendix A).

195 Part II

Radiography demonstrated the findings as are given in Table 3. Subtle new bone formation in the vertebrae was noted as well as progression of osteophytosis in the supplemented groups, but not in the control group. In the all groups, no abnormalities were found on force plate analysis. There were no significant differences in force plate analysis data between all groups.

Histology and immunohistochemical stainings of the liver biopsies are demonstrated by representative examples in Figure 2. Liver biopsies revealed significantly increased numbers of swollen stellate cells in the supplemented groups compared to the control group (p<0.001), but no significant differences were found between the supplemented groups (Table 4). The K19 expression was significantly higher in the HA group and HAHD group compared to the HAMD group (p=0.003 and p=0.002, respectively), but no significant differences were found between the supplemented groups and the control group (Table 5). There were no significant differences in Ki67 expression between groups. The α-SMA expression was significantly higher in groups HA and HAHD compared to the control group (p=0.007 and p=0.02, respectively), and tended to be higher in the HAMD group compared to the controls (p=0.07). In all supplemented groups SR-staining of the liver parenchyma was significantly higher compared to the control group (p<0.001), with no differences between the supplemented groups (Table 4).

Discussion

12 The increased dietary intake of vitamin A and vitamin D from the supplements in this study was reflected by increased plasma levels of retinol in the HA, HAMD, and HAHD group, and by the increased plasma levels of 25OHD in the HAMD and HAHD group, without signs of vitamin D intoxication since all groups maintained normocalcemia and normophosphatemia. Increased dietary vitamin A intake was also demonstrated by the higher amount of retinol and retinyl esters in the liver biopsies in all supplemented groups compared to the control group. The high retinol level of the liver indicates accumulation of retinoids in the liver, which also occurs in arctic top predators. However, those predators have great storage capacity of their hepatic stellate cells and consequently normal plasma retinol levels, and are thus resistant to hypervitaminosis A (Senoo et al. 2012). Cats can apparently not cope with the amounts of retinol offered in this study, as the plasma levels of retinol significantly in all supplemented groups increased compared to the controls, indicating insufficient storage capacity for these amounts of retinol in the hepatic stellate cells.

In the course of the study, in the supplemented groups radiographic new bone formation in the vertebrae was demonstrated, as well as progression of OA. We cannot exclude that the progression of the osteophytosis in the supplemented groups is due to primary OA due to aging of the cats (Slingerland et al. 2011), however, this was not demonstrated in the control group. High vitamin A intake has been suggested to be a causative factor for secondary OA (Lascelles 2010), which might explain these differences in progression

196 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation of OA between the supplemented groups and the control group. Force plate analysis revealed no significant differences between the supplemented groups and the control group, indicating no painful or functional disturbances of the limbs. The lesions previously demonstrated in cats with hypervitaminosis A are the formation of new bone, caused by the intake of raw liver and/or vitamin A and D supplements. Therefore we hypothesized that, in hypervitaminosis A, excessive vitamin D is a prerequisite for development of these skeletal lesions. However, even in the presence of vitamin D close to the safe upper limit according to NRC 2006, the observed skeletal changes were subtle when compared to the new bone formation demonstrated in cats on a diet containing raw liver (Polizopoulou et al. 2005) or vitamin A and D supplements (Hazewinkel and Goedegebuure 1981). The question remains why the classical skeletal symptoms of hypervitaminosis A (i.e. hyperostosis of the cervical vertebrae and around large joints) did not develop in this study. This may be the result of the lower dosage of vitamin A compared to feeding of raw liver (i.e. beef liver (12,517 IU per 100kcal ME) and pork liver (16,157 IU per 100kcal ME) (USDA food database 2013) containing twice as much vitamin A compared to the amounts used in this study), by the duration of the study, which was 18 months compared to 2-5 years in case reports (Polizopoulou et al. 2005), or by other aspects of raw liver feeding. A low calcium intake together with high vitamin D intake may increase osteoclast activity and may be a prerequisite to develop periostal new bone formation (Ryan et al. 2013). Feeding a raw liver diet with its low calcium content (Beef liver: 0.5mg/kg; Pork liver 0.9mg/kg) (USDA food database 2013) compared to the standard diet may therefore activate bone metabolism. 12 The most significant pathology due to high vitamin A intake in this study was early, subclinical liver fibrosis. In the liver biopsies an increased number of swollen stellate cells was demonstrated in the HE staining in all vitamin A supplemented groups, indicating storage of retinol (Senoo et al. 2012). These stellate cells were also activated as can be concluded by the increased α-SMA expression (Kawada et al. 1992). The SR staining revealed formation of fibrous tissue within the liver in the vitamin A supplemented groups (Saxena et al. 2002).

This is the first study reporting liver pathology in cats on high vitamin A intake and is similar to liver pathology in hypervitaminosis A in humans. Activation of hepatic stellate cells due to overload of vitamin A and leading to an up-regulation of transforming growth factor- β may drive this process (Nollevaux et al. 2006, Mallat and Lotersztajn 2013). How high vitamin A intake leads to activation of hepatic stellate cells has not been elucidated, but might be the result of vitamin A accumulation within the hepatic stellate cells exceeding the retinol binding protein capacity causing the release of free retinol and altered gene expression. The accumulation of retinoids can also enhance formation of polar metabolites causing oxidative stress, followed by necrosis of the surrounding hepatocytes which in turn activates hepatic stellate cells to become pro-fibrotic (Nollevaux et al. 2006). Interestingly, in the HAMD group a lower α-SMA expression and lower K19 staining of the liver biopsies was found compared to the HAHD and HA group, indicating

197 Part II

less activation of hepatic stellate cells, and a lower need for hepatocyte regeneration, respectively. We postulate that a moderate level of vitamin D supplementation is protective for development of vitamin A induced liver fibrosis in the cats in this study. The anti-fibrotic effect of vitamin D was also demonstrated in two different mice models of renal disease (Ito et al. 2013). This could be explained by a blockage of the transforming growth factor-β-SMAD signaling pathway, which is known to regulate fibrosis-associated gene expression (Bonventre 2013), as was demonstrated in cell culture of hepatic stellate cells of rats and mice (Shah et al. 2013). The fact that the HAHD group did not show this protective effect of vitamin D on vitamin A induced liver fibrosis, suggests that there is an optimum level of vitamin D supplementation. This optimum in vitamin D as a protective agent for vitamin A induced pathology is in agreement with previous findings in rats (Moore and Wang 1945) and is probably due to insufficient 24-hydroxylase activity in case of excess vitamin D intake. The latter results in an increase in plasma 25OHD which also has affinity for the vitamin D receptor and prevents blockage of the transforming growth factor-β-SMAD signaling pathway by 1,25-dihydroxyvitamin D (Jones 2008), thereby reducing the anti-fibrotic effect of 1,25-dihydroxyvitamin D (Potter et al. 2013).

Conclusions

Cats consuming vitamin A at the current upper limit (NRC 2006) neither showed abnormal locomotion nor clinical symptoms of liver failure after 18 months of supplementation. However, the subtle skeletal changes and the liver pathology in these cats indicate that 12 current vitamin A NRC recommendations for cats are too high. The addition of vitamin D does not seem to influence skeletal pathology in cats during the 18 month study period. A 5-fold increased dietary vitamin D intake may be protective for development of liver pathology in cats consuming 100 times as much vitamin A, but 65-fold increased dietary vitamin D intake is no longer protective.

Conflict of interest statement

None of the authors has any financial or personal relationships that could inappropriately influence or bias the content of the paper.

Acknowledgements

The authors wish to thank Mrs. I.I.M. van Duiven for taking care of the animals and Mr. A. Doornenbal for his technical assistance. Dr. H.C.M. Heuven is acknowledged for his advises in the statistical analysis. Prof. Dr. J. Rothuizen is acknowledged for taking the liver biopsies.

198 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

Appendix A: Supplementary material

Supplementary data associated with this article can be found, in the online version, at doi: .

Plasma levels of retinol, retinyl esters and 25-hydroxyvitamin D (25OHD) in cats expressed in nmol/L

Retinol Retinyl esters 25OHD 0 6 12 18 0 6 12 18 0 6 12 18 CTR 562±83 645±145A 759±94A 762±124A 1093±723 733±431 347±203 1102±310 122±34 122±46A 129±48A 136±50A HA 736±59 1061±431B 1115±254B 1422±352B 3991±1262 2113±1106 1358±908 3101±1551 169±31 163±64A 127±48A 135±54A HAMD 697±216 1055±233B 1456±384B 1261±223B 1825±566 1932±1144 1797±969 2969±608 109±28 140±50A 175±32B 207±42B HAHD 807±157 1106±216B 1252±213B 1593±317B 1507±1021 1682±1286 1709±606 4775±3966 129±51 357±182B 627±137C 656±96C

A = significantly different from B and C (p<0.05)

199 Part II

References • Hu, L., Lind, T., Sundqvist, A., Jacobson, A., Melhus, H., 2010. Retinoic acid increases • Becker, N., Kienzle, E., 2013. Vitamin A proliferation of human osteoclast progenitors excess by feeding with horse meat products and inhibits RANKL-stimulated osteoclast containing high levels of liver. Tierarztliche differentiation by suppressing RANK. PLoS Praxis Ausgabe K: Kleintiere - Heimtiere 41, One 5, e13305. 31-36. • Ito, I., Waku, T., Aoki, M., Abe, R., Nagai, • Bonventre, J.V., 2013. Antifibrotic vitamin D Y., Watanabe, T., Nakajima, Y., Ohkido, I., analogs. Journal of Clinical Investigation 123, Yokoyama, K., Miyachi, H., Shimizu, T., 4570-4573. Murayama, A., Kishimoto, H., Nagasawa, K., • Cho, D.Y., Frey, R.A., Guffy, M.M., Leipold, Yanagisawa, J., 2013. A nonclassical vitamin H.W., 1975. Hypervitaminosis A in the dog. D receptor pathway suppresses renal fibrosis. American Journal of Veterinary Research 36, Journal of Clinal Investigation 123, 4579– 1597-1603. 4594. • Christi, G.A., 1957. Osteoperiostitis difusa • Jimenez, M.J., Balbin, M., Alvarez, J., Komori, T., anquilosant en el gato. Anales de la facultad Bianco, P., Holmbeck, K., Birkedal-Hansen, H., de veterinaria Uruguay, Montevideo 6, 95-106. López, J.M., López-Otín, C., 2001. A regulatory • Corbee, R.J., Maas, H., Doornenbal, A., cascade involving retinoic acid, Cbfa1, and Hazewinkel, H.A.W., 2014. Front and hind matrix metalloproteinases is coupled to the leg ground reaction forces of walking cats; development of a process of perichondrial assessment and comparison with walking invasion and osteogenic differentiation dogs. The Veterinary Journal. Article in Press. during bone formation. Journal of Cell 12 • DeLuca, H.F.,1982. Metabolism and Biology 155, 1333– 1344. mechanism of action of vitamin D Bone • Jones, G., 2008. Pharmacokinetics of vitamin and Mineral Research Annual 1 William A. D toxicity. American Journal of Clinical Peck (Ed). Excerpta Medica, Amsterdam, The Nutrition 88, 582S-586S. Netherlands. 12-13. • Kawada, N., Klein, H., Decker, K., 1992. • Franch, J., Pastor, J., Franch, B., Durall, I., Eicosanoid-mediated contractility of Manzanares, M.C., 2000. Back-scattered hepatic stellate cells. Biochemical Journal 285, electron imaging of a non-vertebral case of 367-371. hypervitaminosis A in a cat. Journal of Feline • Lascelles, B.D.X., 2010. Feline degenerative Medicine and Surgery 2, 49-56. joint disease. Veterinary Surgery 39, 2-13. • Freytag, T.L., Liu, S.M., Rogers, Q.R., Morris, J.G., • Lind, T., Hu, L., Lind, P.M., Sugars, R., 2003. Teratogenic effects of chronic ingestion Andersson, G., Jacobson, A., Melhus, H., 2012. of high levels of vitamin A in cats. Journal of Microarray profiling of diaphyseal bone of rats Animal Physiology and Animal Nutrition 87, suffering from hypervitaminosis A. Calcified 42-51. tissue international 90, 219-229. • Goedegebuure, S.A., Hazewinkel, H.A., 1981. • Mallat, A., Lotersztajn, S., 2013. Cellular Nutrition and bone metabolism. Tijdschrift Mechanisms of Tissue Fibrosis. 5. Novel voor Diergeneeskunde 106, 234-242. insights into liver fibrosis. American Journal of Physiology - Cell Physiology 305, C789-799. • Moore, T., Wang, Y.L., 1945. Hypervitaminosis A. Biochemical Journal 39, 222-228.

200 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

• Nakane, K., Hoshino, K., Kameyama, • Ryan, J.W., Reinke, D., Kogawa, M., Turner, Y., 1978. Effects of hypervitaminosis A A.G., Atkins, G.J., Anderson, P.H., Morris, H.A., on the development of eyes in genetic 2013. Novel targets of vitamin D activity microphthalmia (mic) in mice. Teratology in bone: Action of the vitamin D receptor 18, 143. in osteoblasts, osteocytes and osteoclasts. • Nollevaux, M.C. , Guiot, Y., Horsmans, Y., Current Drug Targets 14, 1683-1688. Leclercq, I., Rahier, J., Geubel, A.P., Sempoux, • Saxena, N.K., Ikeda, K., Rockey, D.C., C., 2006. Hypervitaminsosis A-induced liver Friedman, S.L., Anania, F.A., 2002. Leptin fibrosis: Stellate cell activation and daily dose in hepatic fibrosis: Evidence for increased consumption. Liver International 26, 182-186. collagen production in stellate cells and lean • NRC, 2006: Nutrient Requirements of Dogs littermates of ob/ob mice. Hepatology 35, and Cats. In: National Academy Press, 762-771. Washington DC. • Seawright, A.A., English, P.B., Gartner, R.J.W., • Pasquinelli, F.S., Botrel, T.E.A., Pasquinelli, 1967. Hypervitaminosis A and deforming R.S., Souza, P.F., Pasquinelli Filho, F., 2007. cervical spondylosis of the cat. Journal of Macroscopic ocular alterations of fetuses of Comparative Pathology 77, 29-39. hypervitaminotic A rats. Revista Brasileira de • Senoo, H., Imai, K., Mezaki, Y., Miura, M., Medicina 64, 161-163. Morii, M., Fujiwara, M., Blomhoff, R., 2012. • Polizopoulou, Z.S., Kazakos, G., Patsikas, M.N., Accumulation of Vitamin A in the Hepatic Roubies, N., 2005. Hypervitaminosis A in the Stellate Cell of Arctic Top Predators. cat: a case report and review of the literature. Anatomical Record 295, 1660-1668. Journal of Feline Medicine and Surgery 7, • Shah, R., Reyes-Gordillo, K., Arellanes-Robledo, 363-368. J., Lechuga, C.G., Hernández-Nazara, Z., Cotty, 12 • Potter, J.J., Liu, X., Koteish, A., Mezey, E., 2013. A., Rojkind, M., Lakshman, M.R., 2013. TGF-β1 1,25-dihydroxyvitamin D3 and its nuclear Up-Regulates the Expression of PDGF-β receptor repress human α1(I) collagen Receptor mRNA and Induces a Delayed PI3K-, expression and type I collagen formation. AKT-, and p70S6K -Dependent Proliferative Liver Pathobiology 33, 677-686. Response in Activated Hepatic Stellate • Reiner, G., Hertrampf, B., Köhler, K., 2004. Cells. Alcoholism: Clinical and Experimental Vitamin A-intoxication in the pig. Tierarztliche Research 37, 1838-1848. Praxis Ausgabe G: Grosstiere - Nutztiere 32, • Shirakami, Y., Lee, S-A., Clugston, R.D., Blaner, 218-224. W.S., 2012. Hepatic metabolism of retinoids • Rothenberg, A.B., Berdon, W.E., Woodard, and disease associations. Biochemica et J.C., Cowles, R.A., 2007. Hypervitaminosis Biophysica Acta 1821, 124-136. A-induced premature closure of epiphyses • Silverman, A.K., Ellis, C.N., Voorhees, J.J., 1987. (physeal obliteration) in humans and calves Hypervitaminosis A syndrome: A paradigm of (hyena disease): A historical review of the retinoid side effects. Journal of the American human and veterinary literature. Pediatric Academy of Dermatology 16, 1027-1039. Radiology 37, 1264-1267. • Slingerland, L.I., Hazewinkel, H.A.W., Meij, B.P., • Rothuizen, J., Twedt, D.C., 2009. Liver Biopsy Picavet, P., Voorhout, G., 2011. Cross-sectional Techniques. Veterinary Clinics of North study of the prevalence and clinical features America: Small Animal Practice 39, 469–480. of osteoarthritis in 100 cats. The Veterinary Journal 187, 304-309.

201 Part II

• Stickel, F., Kessebohm, K., Weimann, R., Seitz, H.K., 2011. Review of liver injury associated with dietary supplements. Liver International 31, 595-605. • Testerink, N., Ajat, M., Houweling, M., Brouwers, J.F., Pully, V.V., van Manen, H., Otto, C., Helms, J., Vaandrager, A.B., 2012. Replacement of retinyl esters by polyunsaturated triacylglycerol species in lipid droplets of hepatic stellate cells during activation, PLoS ONE, 7, e34945. • Wendling, D., Hafsaoui, C., Laurain, J.-M., Runge, M., Magy-Bertrand, N., Prati, C., 2009. Dysphagia and hypervitaminosis A: Cervical hyperostosis. Joint Bone Spine 76, 409-411.

202 Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation

203 204 Chapter 13

Rubens 1639 The Three Graces General discussion

205 13

206 General discussion

Many nutrients are required for the normal development and maintenance of the skeleton. Bone consists of an organic matrix, made of collagen fibers and glycoproteins, and anorganic hydroxyapatite composed of calcium, phosphorus, and water. The organic matter of the bone is being remodeled by bone forming osteoblasts and bone resorbing osteoclasts. Both vitamin A and vitamin D influence osteoblasts and osteoclasts. Imbalances in these nutrients can cause skeletal deformities. Optimal nutrition for skeletal growth in young animals and for skeletal health in adult animals is therefore all about balance.

Obesity is the most prevalent nutritional problem; an energy imbalance due to overnutrition that affects skeletal health. In the current western world of abundance, dogs and cats do not need to scavenge for food or overcome periods of starvation anymore. However, during evolution, dogs and cats (and humans) were selected for efficient use of energy sources (Prentice et al. 2005). Easy keepers had an evolutionary benefit, which now contributes to the obesity epidemic in the western world. Obesity has several effects on health. The most important effects are decreased life span, decreased quality of life, type II diabetes in cats and humans, and osteoarthritis (OA) in dogs and humans (Chapter 2, Osto et al. 2013). In the 18th and 19th century, obesity was considered a show off of welfare, and even was considered as beautiful or ideal. Although in humans we now consider obesity as a serious health issue, in companion animals we still consider obesity or overweight as normal, beautiful, or ideal. This is demonstrated by a recent study that demonstrated a prevalence of overweight/obesity in show dogs of 18.6% (Corbee et al. 2013/Chapter 3). Some popular breeds in the overweight group were Bernese Mountain dogs and Labrador retrievers. In show cats, the situation is even more alarming as a recent study demonstrated a prevalence of overweight/obesity of 13 45.5% (Corbee et al. 2014/Chapter 4). Some popular breeds in the overweight group were British shorthair and Norwegian Forest cats. Judges on dog shows and cats shows, as well as dog and cat breeders should be informed about the consequences of overweight conditions in companion animals and instructed to select dogs and cats with healthy body condition as prize winners, rather than overweight animals. However, nutrition is also about emotion. Owners are reluctant to cut down the amount of food to achieve weight loss in companion animals, and prefer other strategies (i.e. increasing activity). Feeding of animals plays an important role in the human-animal bond and cutting down the amount of food is by the owners regarded as punishment for the animal (Corbee and Endenburg 2012). Obese and/or overweight conditions imply several health issues as was demonstrated in a unique longitudinal study in dogs. Labrador siblings were allocated in 2 groups of 24 dogs each; 1 group was fed ad libitum and 1 group was fed 75% of the amount that the other group had eaten, starting at 8-weeks of age. In the ad libitum fed group, the human end point of the study was reached almost 2 years earlier, the incidence of OA in a diversity of joints was higher (19/24 versus 16/24), and the mean age on which treatment for OA was necessary was about 3 years earlier compared to the restricted group (Lawler et al. 2008). From this study it can be concluded that overweight results in aggravation of OA (Hazewinkel and Corbee 2011/

207 Chapter 2). In overweight conditions, blood flow in adipose tissue is limited. As a result, with increasing amounts of fat stored in the adipose tissue, there is an accumulation of waste products and a decrease in nutrients, which triggers an inflammatory response to improve blood flow. The pro-inflammatory cytokines released by the adipose tissue of overweight individuals can be found in plasma, as well as in the synovial fluid (Lübbeke et al. 2013). Furthermore, leptin levels of plasma and synovial fluid increase as a result of increased size of individual adipocytes, stimulating chondrocytes to produce and release matrix metallo-proteinases (MMPs) capable of degrading extracellular matrix proteins present in cartilage and bone (Vuolteenaho et al. 2014). The release of pro-inflammatory cytokines and leptin by adipose tissue coincides with excessive body weight, the latter causing more stress on the affected joints.

OA can also positively be affected by nutritional support. Weight loss has demonstrated to be effective in dogs suffering from OA (Marshall et al. 2010). During weight loss, the production of leptin and pro-inflammatory cytokines in adipose tissue decreases, resulting in decreased production of MMPs, thus preventing further breakdown of articular cartilage. Furthermore, the loss of body weight results in less weight bearing on the joints. Achieving weight loss in companion animals is extremely difficult for the owners, as even in optimal treatment conditions, long-term success rates of weight loss programs are low (German et al. 2012). The latter is explained by the following: leptin levels are chronically increased in obese individuals resulting into leptin insensitivity. When an animal is leptin insensitive, it will not be satiated, and therefore constantly beg for food (Chapter 2), which puts pressure on the human-animal bond. Prevention and/ or early management of obesity is therefore the key to success. Early intervention will 13 restore the normal balance between leptin and adiponectin in adipose tissue. Future studies on candidate genes to prevent obesity, like melanocortin-4 receptor (mcr4r), should be encouraged to be able to select for non-obese genotypes. Activation of the mc4r plays a central role in the energy balance: mc4r agonists decrease energy intake together with increased energy expenditure (Adan et al. 2006).

An important risk factor for obesity in dogs and cats is neutering, which was also demonstrated in show cats (Corbee et al. 2014/Chapter 4). Because of the altered metabolism and decreased leptin sensitivity, neutered animals have lowered energy requirements (Alexander et al. 2011, Mauvais-Jarvis et al. 2013) and are less easily satiated. Veterinarians should take their responsibilities and give proper nutritional recommendations after neutering.

Treatment of OA with dietary long chain polyunsaturated fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) resulted in improvement of locomotion in dogs (Roush et al. 2010) and cats (Corbee et al. 2013/Chapter 5). The study on the effects of EPA+DHA in dogs with OA used force plate analysis (FPA) as an objective outcome parameter (Roush et al. 2010), while the study in cats with OA used owner’s perception as a more subjective outcome parameter (Corbee et al. 2013/Chapter 5). Future studies

208 General discussion

hunger satiety leptin insensitivity

adiponectin leptin leptin

TNFα IL1 MAF VEGF overweight obesity

Fig. 1: Secretion of hormones in adipose tissue with progressive obesity. TNFα = tumor necrosis factor alpha, IL1 = interleukin-1, MAF = macrophage activating factor, VEGF = vascular endothelial growth factor. Circles represent adipocytes (and their size)

AA (C20) EPA (C20) DHA (C22) COX-2/ Aspirin 15-LO COX-2 5-LO 12/15-LO or P450 12/15-LO 18R-HEPE 17S-HpDHA

PGH2 LTA4 15S-HETE 15R-HETE 5-LO 14S-HpDHA

15-LO 5-LO 5-LO 5Hp-18R-HEPE 17S-HDHA

PGD2 PGE2 LTC4 LTB4 LXA4 AT-LXA4 RvE1 RvE2 RvD1 RvD2 PD1 MaR1

13 Receptors for Rosolution:Resolution: ALX/FRL2, CysLT1, GPR32, BLT1, CMLKLR1, ??????

Acute In ammation Resolution of In ammation

Failure of Stop PMN recruitment Block LTs Monocyte recruitment Non-phlogistic Resolution and PG Reduce cytokine release removal of apoptotic PMN

Chronic In ammation Chronic In ammation

Fig. 2: Eicosanoids derived from arachidonic acid (AA), eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA), and their role in the inflammatory response

on effects of EPA+DHA in cats with OA using FPA as outcome parameter (Corbee et al. 2014/Chapter 11) should be performed to have a higher grade of evidence to support the use of EPA+DHA for OA in cats. So far, only one study has demonstrated effects of nutrition on OA in cats using FPA as an outcome parameter (Lascelles et al. 2012). EPA competes with arachidonic acid (AA), and therefore addition of EPA to the diet results in the production of less anti-inflammatory cytokines (including Prostaglandins of the 3-series and Leukotrienes of the 5-series) compared to the pro-inflammatory cytokines derived from AA (Prostaglandins of the 2-series and Leukotrienes of the 4-series) as

209 revealed from the study in cats (Chapter 5). DHA is the precursor of several resolvins that are required for termination of the inflammatory response. Addition of DHA to the diet will aid in preventing an acute inflammatory response to become chronic. Both the quantity and the optimal ratio between EPA:DHA are still debatable, a recent review gives recommendations for the quantity of EPA+DHA for different indications, and at least 60% of total EPA+DHA as EPA is currently recommended (Bauer 2011). Optimal dosage for EPA+DHA still needs to be determined. The amount of EPA+DHA suggested to be effective in dogs with OA are 310mg/kg0.75 body weight (Bauer 2011) and for cats 146mg/ kg body weight (Corbee et al. 2013/Chapter 5). When using high dosages of EPA+DHA, side effects can be seen (e.g. fishy odor of the skin, diarrhea, delayed wound healing, delayed blood clotting, weight gain) (Lenox and Bauer 2013).

The effects of nutrition on skeletal health has been evident for many years, and played a role in the discovery of vitamin A and vitamin D. In man, rickets (vitamin D deficiency) was already described by F. Glisson in 1650 (Wolf 2004). Sunlight has long been recognized as a major provider of vitamin D for humans; radiation in the ultraviolet B (UVB) (290–315 nm) portion of the solar spectrum photolyzes 7-dehydrocholesterol (7-DHC) in the skin to pre-vitamin D3, which, in turn, is converted by a thermal process to vitamin D3 (Webb et al. 1988). The “sunshine vitamin” is however not synthesized in sufficient amount in carnivorous species, as was demonstrated by low levels of the vitamin D precursor 7-DHC in skin samples (Corbee et al. 2014/Chapter 7). The 7-DHC concentration in the skin of carnivores found in this study is 4-40 times lower compared to skin of sheep and goats (Kohler et al. 2013). Although 7-DHC content of skin is regarded as an indicator of vitamin D synthesizing capacity, definitive conclusions on vitamin D synthesizing capacity can 13 only be made when in vivo studies are performed. In dogs (Hazewinkel and Tryfonidou 2002) and cats (Morris 1999) these studies have been performed and demonstrated insufficient vitamin D synthesizing capacity. Dogs and cats therefore solely depend on dietary vitamin D intake to meet their vitamin D requirements, which is probably also true for the carnivorous species studied in Chapter 7. In puppies and kittens, bitch milk and queen milk is the most important source of vitamin D. Because some dogs start to be raised on puppy milk replacers (PMRs) (62% of the breeders reported to use PMRs in our study), the amount of vitamin D in the most popular PMRs were investigated (Corbee et al. 2012/Chapter 8). An important result was that the vitamin D content of PMRs is often higher (on average 35% more in the 8 investigated PMRs in our study) compared to the declaration on the package (in 7 out of 8 investigated PMRs in our study), and sometimes reaches the amount proven to cause disturbances in endochondral ossification when given from 6-21 weeks of age (Tryfonidou et al. 2002). Furthermore, most breeders overfeed their puppies with PMR, which resulted in calcium and vitamin D intakes exceeding the safe upper limit according to NRC 2006. The increased calcium intake before weaning may result in enostosis during the rapid growth phase later in life (Schoenmakers, 1998). The composition of PMRs was compared with the composition of bitch milk and revealed lower protein content, lower fat content, and lower energy density compared to bitch milk. Interestingly, the ratio of 25-hydroxyvitamin

210 General discussion

D (25OHD) to vitamin D in bitch milk is higher compared to the 25OHD to vitamin D ratio in human milk, which might indicate a lower or insufficient hepatic 25-hydroxylase activity in puppies compared to human infants (Ala-Houhala et al. 1988, Corbee et al. 2012/Chapter 8). Further investigations are needed to determine the consequences of this finding for the production of balanced PMRs. The main function of vitamin D in calcium metabolism is to release calcium and phosphorus from the diet, reabsorb calcium and phosphorous from the pre-urine, and stimulate the osteoclasia of mineralized bone, and eventually deposit this available calcium and phosphorus in the newly formed osteoid and cartilage. In puppies suffering from medial coronoid disease (MCD) poor mineralization and/or delayed endochondral ossification has recently been demonstrated (Lau et al. 2013). We hypothesized that this could be the result of vitamin D deficiency, however in Chapter 9 we demonstrated that supplementing the daily ration with vitamin D did not result in a decreased prevalence of MCD, suggesting that other factors play a role in MCD development (Corbee et al. 2014/ Chapter 9). Moreover, additional vitamin D in this study slowed down the endochondral ossification as was noted by more irregular growth plates and increased collagen X staining of the growth plates compared to the control group, in line with a previous report in large breed dogs (Tryfonidou et al. 2003a).

Vitamin D needs to be metabolized to its active metabolites to exert effects. The conversion of 25OHD to the active metabolite 1,25-dihydroxyvitamin D (1,25DHCC) is stimulated by growth hormone (GH), which is mainly produced by the pituitary gland, and insulin-like growth factor I (IGF-I) (Tryfonidou et al. 2003b). It is also known that in growing large breed dogs, plasma levels of GH, IGF-I, and consequently 1,25DHCC are higher compared to small breed dogs (Nap et al. 1993, Tryfonidou et al. 2003b). 13 In adult dogs, a common disease is pituitary dependent hyperadrenocorticism. In humans corticoid treatment resulted in a decrease of plasma levels of GH, IGF-I, and 1,25DHCC in a dose dependent manner, and therefore, hyperadrenocorticism is regarded as an important form of secondary osteopenia in humans (Klaus et al. 2000, Suzuki et al. 1983). Therefore, it was hypothesized that glucocorticoid overload due to pituitary dependent hyperadrenocorticism would result in decreased plasma 1,25DHCC levels in dogs. Furthermore, as a standard of care treatment of pituitary dependent hyperadrenocorticism in dogs is the removal of the pituitary gland, the main source of GH production, we hypothesized that hypophysectomy would result in decreased levels of GH, IGF-I, and 1,25DHCC plasma levels, and consequently increased parathyroid hormone (PTH) secretion. The increase of PTH secretion in dogs has been associated with decreased quality of life, appetite, activity, strength, and life span (Nagode et al. 1996). Interestingly, in contrast to humans, plasma levels of 1,25DHCC remained stable in dogs with pituitary dependent hyperadrenocorticism after hypophysectomy and where comparable to matched controls, despite a significant decrease in plasma GH and IGF-I levels (Corbee et al. 2012/Chapter 10). In young growing dogs, GH and IGF-I are important for stimulation of 1-alpha hydroxylation of 25OHD (Tryfonidou et al. 2003b), however in adult dogs it is obviously not (Corbee et al. 2012/Chapter 10). This leaves PTH as the main

211 factor to stimulate 1-alpha hydroxylase, however its activity is not increased in older dogs on a balanced diet and with normal kidney function and even not increased in dogs before or after treatment of pituitary dependent hyperadrenocorticism. This is probably due to the much lower bone turn-over in adult dogs than in adult man.

Interestingly, the glucocorticoid excess during hyperadrenocorticism coincided with a significant decrease in GH plasma levels, however this did not lead to a significant decrease of IGF-I plasma levels. From the study described in Chapter 10 it can be concluded that vitamin D metabolite supplementation is not needed in adult dogs with hyperadrenocorticism before and after hypophysectomy to prevent hyperparathyroidism (Corbee et al. 2012/Chapter 10).

Vitamin A also plays a role in maintaining skeletal health. Vitamin A stimulates cellular differentiation and is required for the activation of osteoclasts. Vitamin A also stimulates 24-hydroxylase, which is required for the formation of 24,25-dihydroxyvitamin D and catabolism of 1,25DHCC into 1,24,25-trihydroxyvitamin D. Vitamin A excess in cats has been studied in the 1960s and 1970s and demonstrated skeletal hyperostosis, mostly in the cervical spine (Seawright et al. 1967, Clark et al. 1970, Seawright et al. 1970, Clark 1973, Seawright and Hrdlicka 1974). Since then, the nutritional recommendations for vitamin A in cats have never been changed (NRC 2006, FEDIAF 2013, AAFCO 2013). In humans, hypervitaminosis A is characterized by liver fibrosis, which was very recently also described in an 8-year old cat, that was fed a raw liver diet and presented skeletal hyperostosis. (Guerra et al. 2014). Subtle bone and liver pathology was also observed in cats when they were fed vitamin A 13 for 18 months at the safe upper limit according to NRC 2006 (Corbee et al. 2014/ Chapter 12). To demonstrate the clinical relevance of the radiological demonstrated bony changes, we studied an objective outcome parameter for evaluation of lameness in dogs for applicability in cats (i.e. FPA) (Corbee et al. 2014/Chapter 11). The subtle bony changes demonstrated in cats that were fed high amounts of vitamin A did not (yet) result in clinical lameness according to FPA. An inquiry was not performed in these cats, as these were research cats that can’t be observed in a household situation, as is required for the commonly used inquiries evaluating lameness in cats (Slingerland et al. 2011, Zamprogno et al. 2010). The clinical relevance of the hepatic fibrosis demonstrated by immunohistochemistry are debatable, as none of the cats demonstrated signs of liver failure. However, we cannot exclude that increased vitamin A intake may indeed by related to early pre-clinical liver pathology.

On the basis that vitamin A and D interact (Chapter 6), the second aim of the study on hypervitaminosis A in cats was to demonstrate if vitamin D supplementation is of influence on the pathology seen in hypervitaminosis A in cats. The classical symptoms of hypervitaminosis A were not observed in this study, not even in the presence of excess vitamin D. In bone, 1,25DHCC enhances vitamin A induced osteoclasia, which results in weakened bone. By traction on the periost by tendons and muscles, this results in

212 General discussion

the formation of reactive bone, which is enhanced by vitamin A induced 24,25DHCC. Although additional vitamin D did not influence bone pathology, it was protective for liver pathology. In the liver, 1,25DHCC inhibits vitamin A induced fibrosis by a blockage of the transforming growth factor-β-SMAD signaling pathway (Bonventre 2013). Interestingly, higher levels of vitamin D were no longer protective (Corbee et al. 2014/ Chapter 12) indicating the importance of a balanced intake of vitamin A and vitamin D. Further research is needed to elucidate the pathophysiology of the bone lesions seen in hypervitaminosis A in cats which is also seen in humans (Wendling et al. 2009). Because most cases of hypervitaminosis A in cats were associated with the chronic consumption of raw liver, it might be that chronicity or other aspects of raw liver influence development of this lesions, like a calcium deficiency or vitamin E deficiency.

213 References • Guerra, J.M., Daniel, A.G.T., Aloia, T.P.A., de Siqueira, A., Fukushima, A.R., Simões, • Adan, R.A.H., Tiesjema, B., Hillebrand, J.J.G., D.M.N., Reche-Júior, A., Cogliati, B., 2014. Fleur, S.E. la, Kas, M.J.H., Krom, M. de, 2006. Hypervitaminosis A-induced hepatic fibrosis The MC4 receptor and control of appetite. in a cat. Journal of Feline Medicine and British Journal of Pharmacology 149, 815–827. Surgery 16, 243-248. • Ala-Houhala, M., Koskinen, T., Parviainen, • Hazewinkel, H.A.W , Tryfonidou, M.A, 2002. M.T., Visakorpi, J.K., 1988. 25-Hydroxyvitamin Vitamin D3 metabolism in dogs. Molecular D and vitamin D in human milk: Effects of and Cellular Endocrinology 197, 23-33. supplementation and season. American • Klaus, G., Jux, C., Fernandez, P., Rodriguez, Journal of Clinical Nutrition 48, 1057-1060. J., Himmele, R., Mehls, O., 2000. Suppression • Alexander, L.G., Salt, C., Thomas, G., of growth plate chondrocyte proliferation Butterwick, R., 2011. Effects of neutering by corticosteroids. Pediatric Nephrology 14, on food intake, body weight and body 612-615. composition in growing female kittens. British • Kohler, M., Leiber, F., Willems, H., Merbold, L., Journal of Nutrition 106, S19-S23. Liesegang, A,, 2013. Influence of altitude on • Bauer, J.E., 2011. Therapeutic use of fish oils in vitamin D and bone metabolism of lactating companion animals. Journal of the American sheep and goats. Journal of Animal Science Veterinary Medical Association 239, 1441- 91, 5259-5268. 1451. • Lascelles, B.D.X., DePuy, V., Thomson, A., • Bonventre, J.V., 2013. Antifibrotic vitamin D Hansen, B., Marcellin-Little, D.J., Biourge, V., analogs. Journal of Clinical Investigation Bauer, J.E., 2010. Evaluation of a therapeutic 123, 4570-4573. diet for feline degenerative joint disease. • Clark, L., 1973. Growth rates of epiphyseal Journal of Veterinary Internal Medicine 24, 13 plates in normal kittens and in kittens fed 487-495. excess vitamin A. A fluorescent labelling and • Lau, S.F., Hazewinkel, H.A.W., Grinwis, histological study. Journal of Comparative G.C.M., Wolschrijn, C.F., Siebelt, M., Vernooij, Pathology 83, 447-460. J.C.M., Voorhout, G., Tryfonidou, M.A., • Clark, L., Seawright, A.A., Gartner, R.J.W., 2013b. Delayed endochondral ossification 1970. Longbone abnormalities in kittens in early medial coronoid disease (MCD): A following vitamin a administration. Journal of morphological and immunohistochemical Comparative Pathology 80, 113-120. evaluation in growing Labrador retrievers. The • Corbee, R.J., Endenburg, N., 2012. Influence of Veterinary Journal 197, 731–738. obesity on the owner’s perception of welfare • Lawler, D.F., Larson, B.T., Ballam, J.M., Smith, in dogs. Proceedings ESVCN Congress 2012, G.K., Biery, D.N., Evans, R.H., Greeley, E.H., Bydgoszcz, Poland. Segre, M., Stowe, H.D., Kealy, R.D., 2008. Diet • German, A.J., Holden, S.L., Morris, P.J., Biourge, restriction and ageing in the dog: major V., 2012. Long-term follow-up after weight observations over two decades. British management in obese dogs: The role of diet Journal of Nutrition 99, 793-805. in preventing regain. The Veterinary Journal • Lenox, C.E., Bauer, J.E., 2011. Potential adverse 192, 65-70. effects of omega-3 fatty acids in dogs and cats. Journal of Veterinary Internal Medicine 27, 217-226.

214 General discussion

• Lübbeke, A., Finckh, A., Puskas, G.J., Suva, D., • Roush, J.K., Cross, A.R., Renberg, W.C., Dodd, Lädermann, A., Bas, S., Fritschy, D., Gabay, C., C.E., Sixby, K.A., Fritsch, D.A., Allen, T.A., Jewell, Hoffmeyer, P., 2013. Do synovial leptin levels D.E., Richardson, D.C., Leventhal, P.S., Hahn, correlate with pain in end stage arthritis? K.A., 2010. Evaluation of the effects of dietary International Orthopaedics 37, 2071-2079. supplementation with fish oil omega-3 • Marshall, W.G., Hazewinkel, H.A., Mullen, D., fatty acids on weight bearing in dogs with Meyer, G. de, Baert, K., Carmichael, S., 2010. osteoarthritis. Journal of the American The effect of weight loss on lameness in Veterinary Medical Association 236, 67-73. obese dogs with osteoarthritis. Veterinary • Schoenmakers,I., 1998. Thesis: Modulation Research Communications 34, 241-253. of calcium regulation by excessive calcium • Mauvais-Jarvis, F., Clegg, D.J., Hevener, A.L., intake in dogs, ISBN: 90-393-1829-8. 2013. The role of estrogens in control of • Seawright, A.A., English, P.B., Gartner, R.J.W., energy balance and glucose homeostasis. 1967. Hypervitaminosis A and deforming Endocrinology Reviews 34, 309-338. cervical spondylosis of the cat. Journal of • Morris, J.G., 1999. Ineffective vitamin D Comparative Pathology 77, 29-39. synthesis in cats is reversed by an inhibitor of • Seawright, A.A., English, P.B., Gartner, R.J., 7-dehydrocholestrol-δ7-reductase. Journal of 1970. Hypervitaminosis A of the cat. Advances Nutrition 129, 903-908. in veterinary science and comparative • Nagode, L.A., Chew, D.J., Podell, M., medicine 14, 1-27. 1996. Benefits of calcitriol therapy and • Seawright, A.A., Hrdlicka, J., 1974. serum phosphorus control in dogs and Pathogenetic factors in tooth loss in young cats with chronic renal failure. Both are cats on a high daily oral intake of vitamin A. essential to prevent or suppress toxic Australian Veterinary Journal 50, 133-141. hyperparathyroidism. Veterinary Clinics of • Slingerland, L.I., Hazewinkel, H.A.W., Meij, B.P., North America – Small Animal Practice 26, Picavet, P., Voorhout, G., 2011. Cross-sectional 13 1293-1330. study of the prevalence and clinical features • Nap, R.C., Mol, J.A., Hazewinkel, H.A.W., 1993. of osteoarthritis in 100 cats. The Veterinary Age-related plasma concentrations of growth Journal 187, 305-309. hormone (GH) and insulin-like growth factor • Suzuki, Y., Ichikawa, Y., Saito, E., Homma, I(IGF-I) in Great Dane pups fed different M., 1983. Importance of increased urinary dietary levels of protein. Domestic Animal calcium excretion in the development of Endocrinology 10, 237-247. secondary hyperparathyroidism of patients • NRC, 2006. Nutrient Requirements of Dogs under glucocorticoid therapy. Metabolism: and Cats. In: National Academy Press, Clinical and Experimental 32, 151-156. Washington DC. • Tryfonidou, M.A., Holl, M.S., Stevenhagen, • Osto, M., Zini, E., Reusch, C.E., Lutz, T.A., 2013. J.J., Buurman, C.J., Deluca, H.F., Oosterlaken- Diabetes from humans to cats. General and Dijksterhuis, M.A., Van Den Brom, W.E., Comparative Endocrinology 182, 48-53. Van Leeuwen, J.P.T.M., Hazewinkel, H.A.W., • Prentice, A.M., Rayco-Solon, P., Moore, S.E., 2003a. Dietary 135-fold cholecalciferol 2005. Insights from the developing world: supplementation severely disturbs the thrifty genotypes and thrifty phenotypes. endochondral ossification in growing dogs. Proceedings of the Nutrition Society 64, Domestic Animal Endocrinology 24, 265-285. 153-161.

215 • Tryfonidou, M.A., Holl, M.S., Vastenburg, M., Oosterlaken-Dijksterhuis, M.A., Birkenhäger- Frenkel, D.H., Van Den Brom, W.E., Hazewinkel, H.A.W., 2003b. Hormonal regulation of calcium homeostasis in two breeds of dogs during growth at different rates. Journal of Animal Science 81, 1568-1580. • Tryfonidou, M.A., Stevenhagen, J.J., Van Den Bemd, G.J.C.M., Oosterlaken-Dijksterhuis, M.A., Deluca, H.F., Mol, J.A., Van Den Brom, W.E., Van Leeuwen, J.P.T.M., Hazewinkel, H.A.W., 2002. Moderate cholecalciferol supplementation depresses intestinal calcium absorption in growing dogs. Journal of Nutrition 132, 2644-2650. • Vuolteenaho, K. , Koskinen, A., Moilanen, E., 2014. Leptin - A link between obesity and osteoarthritis: Applications for prevention and treatment. Basic and Clinical Pharmacology and Toxicology 114, 103-108. • Webb, A.R., Kline, L., Holick, M.F., 1988. Influence of season and latitude on the cutaneous synthesis of vitamin D3: Exposure to winter sunlight in Boston and Edmonton 13 will not promote vitamin D3 synthesis in human skin. Journal of Clinical Endocrinology and Metabolism 67, 373-378. • Wendling, D., Hafsaoui, C., Laurain, J.M., Runge, M., Magy-Bertrand, N., Prati, C., 2009. Dysphagia and hypervitaminosis A: Cervical hyperostosis. Joint Bone Spine 76, 409-411. • Wolf, G., 2004. The Discovery of Vitamin D: The Contribution of Adolf Windaus. Journal of Nutrition 134, 1299-1302. • Zamprogno, H., Hansen, B.D., Bondell, H.D., Sumrell, A.T., Simpson, W., Robertson, I.D., Brown, J., Pease, A.P., Roe, S.C., Hardie, E.M., Wheeler, S.J., Lascelles, D.X., 2010. Item generation and design testing of a questionnaire to assess degenerative joint disease-associated pain in cats. American Journal of Veterinary Research 71, 1417-1424.

216 General discussion

217 218 Summary

219 In this thesis, the influence of nutrition on skeletal health was studied. This thesis is divided in two parts. Part I focuses on the effects of nutrition on osteoarthritis, whereas in Part II of this thesis, the effects of imbalances of vitamin A and D on skeletal health are described.

Part I

Obesity is a common disease in companion animals. It is caused by an imbalance between energy intake and energy expenditure, and defined as a condition in which excess body fat has accumulated to the extent that it may be harmful to health. In dogs and cats obesity often exacerbates osteoarthritis (OA). In overweight conditions, pro- inflammatory cytokines are released by the adipose tissue and can be found in plasma, as well as in the synovial fluid. Plasma leptin levels also increase in overweight conditions, as it is produced by the adipose tissue when fat accumulates in the adipocytes. Leptin is a strong satiating signal and thus decreases food intake. However, leptin levels not only increase in plasma, but also in synovial fluid, thereby stimulating chondrocytes to produce and release matrix metallo-proteinases (MMPs) capable of degrading extracellular matrix proteins present in cartilage and bone. Furthermore, excessive body weight causes more stress on the affected joints Chapter( 2).

In the 18th and 19th century, obesity was considered a show off of welfare, and even was considered as beautiful or ideal. Although in humans we now consider obesity as a serious health issue, in companion animals obesity or overweight is still considered as normal, beautiful, or ideal as was demonstrated in Chapters 3 and 4. Judges on dog shows and cats shows, as well as dog and cat breeders should be informed about the consequences of overweight conditions in companion animals and instructed to select Sm dogs and cats with healthy body condition as prize winners, rather than overweight animals. An important risk factor for obesity in dogs and cats is neutering, which was also demonstrated in show cats in Chapter 4. Because of the altered metabolism and decreased leptin sensitivity of the satiety center, neutered animals have lowered energy requirements and are less easily satiated. Veterinarians should take their responsibilities and give proper nutritional recommendations after neutering.

In old dogs, OA is a common disease and several treatment options have been investigated and are generally accepted by clinicians. Recently, the high prevalence of OA in elderly cats was identified, and it was clear that these cats demonstrate more subtle behavioral changes (e.g. sleep more, move less, less interaction with the owner, defecating outside the litter box, less grooming), rather than obvious lameness as can be observed in dogs suffering from OA. One of the nutraceuticals used in dogs suffering from OA are the dietary long-chain omega-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). These fatty acids have been proved to be effective in dogs suffering from OA by mediating the inflammatory response. In a double blinded cross-over design study in client-owned cats with naturally occurring OA, we

220 Summary

demonstrated that treatment of these cats with dietary supplementation with EPA and DHA also resulted in improvement of locomotion (Chapter 5).

Part II

In Chapter 6 an overview is given of two important vitamins in skeletal health, i.e. vitamin D and vitamin A. The effects of these vitamins on the organism, with special interest to skeletal health, are presented. Vitamin D is important for skeletal health as it plays a central role in calcium metabolism. The role of vitamin D in calcium metabolism is to release calcium and phosphorus from the diet, to reabsorb calcium and phosphorous from the pre-urine, and to stimulate the osteoclasia of mineralized bone, and eventually deposit this available calcium and phosphorus in the newly formed osteoid and cartilage (Chapter 6).

Sunlight has long been recognized as a major provider of vitamin D for humans. The “sunshine vitamin” is however not synthesized in sufficient amount in the investigated carnivorous species, as was demonstrated by low levels of the vitamin D precursor 7-dehydrocholesterol (7-DHC) in their skin samples (Chapter 7). The investigated species thus are largely dependent on dietary intake of vitamin D.

High calcium intake in puppies from 3-6 weeks (pre-weaning) has been proven to influence skeletal development at older age. An inquiry among 184 German Shepherd breeders was performed to investigate the brands and amounts of used puppy milk replacers (PMRs). Since quite some puppies start to be raised on PMRs, the calcium content of the PMRs was of concern, and investigated in the eight most used PMRs. In the amounts used by most breeders, the total calcium intake may exceed the calcium requirements. The increased calcium intake before weaning may result in enostosis Sm during the rapid growth phase later in life. In addition, the vitamin D content in the eight most popular PMRs was investigated (Chapter 8). An important result was that the vitamin D content of PMRs is often higher compared to the declaration on the package, and sometimes reaches the amount in earlier studies proven to cause disturbances in endochondral ossification when given from 6-21 weeks of age. Furthermore, most breeders overfeed their puppies with PMR, which resulted in vitamin D intakes exceeding the safe upper limit according to NRC 2006. Interestingly, it revealed from these studies that the concentration of 25-hydroxycholecalciferol (25OHD) in bitch’s milk exceeds the cholecalciferol (i.e. genuine vitamin D) content, which may indicate the inability of young puppies to hydroxylate vitamin D which may have serious consequences for their early skeletal development.

Another common developmental orthopedic diseases in young dogs is medial coronoid disease (MCD). Delayed endochondral ossification plays a crucial role in MCD development. This delayed endochondral ossification (i.e. delay in cartilage mineralization and chondrocyte maturation) might be due to a (relative) vitamin D deficiency. The study

221 in Chapter 9 demonstrated that supplementing the daily ration of young growing dogs susceptible to develop MCD with vitamin D did not result in a decreased prevalence of MCD. These findings are suggesting that other factors play a role in delayed endochondral ossification preceding MCD development.

Vitamin D needs to be metabolized to its active metabolites to exert effects. Vitamin D is converted to 25OHD in the liver. The conversion of 25OHD to the active metabolite 1,25-dihydroxyholecalciferol (1,25DHCC) is stimulated by growth hormone (GH), which is mainly produced by the pituitary gland, and insulin-like growth factor I (IGF-I). In dogs with Cushings disease (hyperadrenocorticism) an important source of GH and consequently of IGF-I is suppressed by the negative feedback of cortisol on the pituitary gland. In 85% of the cases of Cushings disease, hyperadrenocorticism is pituitary dependent (thus 15% of the cases are caused by a tumor of the adrenal gland). Treatment options are either removal of the tumor of the adrenal gland, destruction of the adrenal glands, medication to reduce the release of cortisol, or removal of the pituitary gland (e.g. hypophysectomy). The removal of the pituitary gland results in the removal of an important source of GH, and in humans this results in osteoporosis due to decreased 1,25DHCC plasma levels and concomitant increased parathyroid hormone (PTH) secretion. The latter is hold responsible for a decreased quality of life in dogs, and was the rationale for the study presented in Chapter 10. However, in adult dogs, 1,25DHCC and PTH plasma levels were not different before and after hypophysectomy compared to healthy control dogs, probably due to the much lower bone turn-over in adult dogs than in adult man. In conclusion, neither in dogs with Cushings disease nor after hypophysectomy, there is a need for vitamin D supplementation.

Vitamin A is the other vitamin which plays a role in maintaining skeletal health. As Sm reviewed in Chapter 6, vitamin A stimulates cellular differentiation, is required for the activation of (bone resorbing) osteoclasts, and stimulates 24-hydroxylase, which is required for the formation of 24,25-dihydroxycholecalciferol and catabolism of 1,25DHCC into 1,24,25-trihydroxycholecalciferol. Vitamin A excess in cats has been studied in the 1960s and 1970s and demonstrated skeletal hyperostosis, mostly in the cervical spine. Since then, the nutritional recommendations for vitamin A in cats have never been changed. In Chapter 12, subtle bone and liver pathology are described as were observed in a longitudinal, comparative study in 24 cats when fed vitamin A for 18 months at the safe upper limit according to NRC 2006. To demonstrate the clinical relevance of the radiological demonstrated bony changes, we studied an objective outcome parameter available for evaluation of lameness in dogs for applicability in cats (i.e. force plate analysis (FPA)) in Chapter 11. Cats are different from dogs in several force plate characteristics. The high reproducibility and sensitivity of this non-invasive technique forms a valuable tool to perform objective follow-up of locomotion in cats. The subtle bony changes demonstrated in cats fed high amounts of vitamin A did not (yet) result in clinical lameness according to FPA (Chapter 12). The clinical relevance of the hepatic fibrosis demonstrated by immunohistochemistry are debatable, as none of the cats demonstrated signs of

222 Summary

liver failure. However, we cannot exclude that increased vitamin A intake may indeed by related to early pre-clinical liver pathology with serious consequences later in life. It is remarkable that moderate increased vitamin D supplementation (as present in natural feeding stuffs like raw liver) decreases the detrimental influence of excess vitamin A on liver fibrosis, but not on the skeleton, possibly due to the anti-fibrotic effect of moderate vitamin D. The findings as described in Chapter 12, warrants reconsideration of the regulating authorities regarding the vitamin A dietary content in relation to the vitamin D content in cat food.

From the studies described in this thesis it can be concluded that an imbalance in energy intake and energy expenditure, causing frequently observed overweight in dogs and cats, is harmful for joint function, and that increased intake of EPA+DHA in addition to weight reduction can be applied to treat OA. It has been shown that vitamin D in puppy milk replacer might be ineffective. Furthermore, the addition of vitamin D to young growing dogs to prevent fragmented coronoid process is superfluous, and to dogs with hyperadrenocorticism not necessary. However, the addition of a limited amount of vitamin D to cats with excessive vitamin A intake protected against the fibrosis of the liver without noticeable potentiating the osseous effects.

223 224 Samenvatting

225 In dit proefschrift is de invloed van voeding op het skelet van honden en katten bestudeerd. Het proefschrift is opgedeeld in twee delen. Deel I beschrijft de invloed van voeding op osteoartrose. In deel II staan de invloed van vitamine A en vitamin D op het skelet beschreven.

Deel I

Obesitas is een veel voorkomende ziekte bij gezelschapsdieren. Obesitas wordt veroorzaakt door een disbalans tussen energieopname en -verbruik en is een toestand waarbij overmatige ophoping van lichaamsvet leidt tot schadelijke effecten voor de gezondheid. Bij honden en katten verergert obesitas osteoartrose (OA). Bij dieren met overgewicht scheidt het vetweefsel stoffen (pro-inflammatoire cytokines) af die de ontstekingsreactie versterken. Deze cytokines komen via het vetweefsel terecht in de bloedbaan en daarmee ook in de gewrichtsvloeistof. Een andere belangrijke stof uit vetweefsel is het verzadigingshormoon leptine, dat wordt afgegeven wanneer vet opstapelt in vetcellen en zorgt voor een daling van de voedselopname. Bij dieren met overgewicht wordt leptine in grotere hoeveelheden afgegeven. Als te veel leptine in de gewrichtsvloeistof terecht komt, dan stimuleert het de kraakbeencellen tot de aanmaak van matrix metallo-proteinases (MMPs) die zorgen voor de afbraak van de eiwitten die aanwezig zijn in de extracellulaire matrix van kraakbeen en bot. Daarnaast zorgt overgewicht ook voor een verhoogde belasting op de artrotische gewrichten (Hoofdstuk 2).

In de 18e en 19e eeuw, werd obesitas en/of overgewicht in de westerse wereld gezien als een teken van welvaart en werd zelfs gezien als schoonheid of ideaalbeeld. Hoewel we bij mensen nu inzien dat overgewicht kan leiden tot ernstige gezondheidsklachten, beschouwen wij bij een aantal honden en katten rassen overgewicht nog steeds als schoonheid of ideaalbeeld, zoals blijkt uit de studies in Hoofdstukken 3 en 4. Keurmeesters Sm op honden- en kattenshows, alsmede honden- en kattenfokkers dienen goed te worden geïnformeerd over de gevolgen van overgewicht bij gezelschapsdieren en zouden honden en katten met een gezond gewicht moeten aanwijzen als prijswinnaars, in plaats van honden en katten met overgewicht. Castratie blijkt ook een belangrijke risicofactor te zijn voor overgewicht; iets dat danook werd gevonden in de studie bij showkatten in Hoofdstuk 4. Vanwege een lager metabolisme en een verminderde gevoeligheid voor leptine van het verzadigingscentrum hebben gecastreerde dieren een lagere energiebehoefte en zijn ze minder snel verzadigd. Dierenartsen dienen hun verantwoordelijkheid te nemen door deze gevolgen van castratie goed met eigenaren door te spreken en goede voedingsadviezen te geven na castratie om overgewicht op latere leeftijd te voorkomen.

Bij oudere honden is OA een veelvoorkomende ziekte, welke gemakkelijk wordt herkend (bijvoorbeeld kreupelheid, moeilijk uit de mand komen, minder gemakkelijk/niet meer in de auto springen). Meerdere behandelingsmogelijkheden zijn goed onderzocht en

226 Samenvatting

worden veelvuldig toegepast in de praktijk. In een recent onderzoek is aangetoond dat OA ook zeer vaak voorkomt bij oudere katten, echter de symptomen zijn veel subtieler (meer slapen, minder bewegen, minder interactie met de eigenaar, naast de bak poepen, minder goede vachtverzorging). Eén van de voedingsstoffen met bewezen effecten op OA bij honden zijn de meervoudig onverzadigde lang keten omega-3 vetzuren eicosapentaeenzuur (EPA) en docosahexaeenzuur (DHA). Deze vetzuren verminderen de ontstekingsreacties in de gewrichten. In een dubbel-blinde placebo-gecontroleerde cross-over studie is aangetoond dat katten met OA verbeteren wanneer EPA en DHA worden gegeven (Hoofdstuk 5).

Deel II

In Hoofdstuk 6 is een overzicht gegeven van twee belangrijke vitamines die invloed hebben op de skeletgezondheid; vitamine D en vitamine A. De effecten van deze vitamines op het organisme, met de nadruk op het skelet, zijn beschreven. Vitamin D is belangrijk voor het skelet omdat het een centrale rol speelt in de calcium stofwisseling. Vitamine D bevordert de opname van calcium en fosfor vanuit het dieet, de terugresorptie van calcium en fosfor uit de voor-urine, en het vrijmaken van calcium en fosfor vanuit het bot door stimulatie van botafbraak (osteoclasie). Daarnaast stimuleert vitamine D de mineralisatie van nieuw gevormd osteoid (voorloper van bot) en kraakbeen (Hoofdstuk 6).

Ultraviolet B licht uit een deel van het zonlicht spectrum is een belangrijke factor voor de vorming van vitamine D in de huid bij mensen. Van de onderzochte carnivoren is aangetoond dat zij onvoldoende “zonlicht vitamine” kunnen aanmaken in de huid, omdat er lage hoeveelheden van de voorloper 7-dehydrocholesterol (7-DHC) in de huid aanwezig zijn (Hoofdstuk 7).

Eerdere studies hebben aangetoond dat een hoge calciuminname bij pups van de leeftijd Sm 3-6 weken de skeletontwikkeling op latere leeftijd beïnvloedt. Om te onderzoeken welke kunstmelk (puppy melkpoeders = PMRs) worden gebruikt en in welke hoeveelheden, werd een enquête gehouden onder 184 Duitse herder fokkers. 62% van de geïnterviewde Duitse herder fokkers bleek wel eens kunstmelk te gebruiken. Van de 8 meest gebruikte PMRs is het calciumgehalte onderzocht. In de hoeveelheden die worden aangehouden door de meeste fokkers krijgen de pups meer dan de voorgeschreven hoeveelheid calcium binnen. Te hoge calciuminname voor het spenen kan de oorzaak zijn van groeipijn (enostosis) op latere leeftijd. Naast het calciumgehalte is het vitamine D gehalte van de acht meest gebruikte PMRs onderzocht (Hoofdstuk 8). Een belangrijke conclusie was dat het vitamine D gehalte in PMRs vaak hoger is dan op de verpakking staat vermeld. Bovendien overvoeren de meeste fokkers hun pups met PMRs waardoor er meer dan de voorgeschreven hoeveelheid vitamine D wordt opgenomen. Daarbij worden soms hoeveelheden bereikt waarvan in eerdere studies is aangetoond dat deze schadelijk kunnen zijn voor de ontwikkeling van het skelet. Interessant was de bevinding

227 dat de hoeveelheid 25-hydroxycholecalciferol (25OHD) in tevenmelk hoger was dan het cholecalciferol (het gewone vitamine D) gehalte, wat zou kunnen wijzen op het feit dat jonge pups vitamine D niet goed kunnen omzetten in 25OHD. Het geven van cholecalciferol in plaats van 25-OHD in PMRs kan mogelijk grote gevolgen hebben voor de skeletontwikkeling.

Een ander veelvoorkomend skeletprobleem bij jonge honden is het los processus coronoideus medialis (LPC). Vertraagde endochondrale ossificatie (vertraagde mineralisatie van kraakbeen en vertraagde maturatie van condrocyten) speelt een cruciale rol bij het ontstaan van LPC. Deze vertraagde mineralisatie kan het gevolg zijn van een (relatief) vitamine D tekort. In Hoofdstuk 9 is aangetoond dat het geven van extra vitamin D bovenop de dagelijkse voeding geen beschermend effect heeft op de vorming van LPC. Andere factoren die mineralisatie vertragen liggen dus vermoedelijk ten grondslag aan de vorming van LPC.

Vitamin D dient te worden omgezet in actieve metabolieten alvorens het zijn effecten kan uitoefenen. Vitamine D wordt in de lever omgezet tot 25OHD. De omzetting van 25OHD in de nier naar de actieve metaboliet 1,25-dihydroxycholecalciferol (1,25DHCC) wordt gestimuleerd door groei hormoon (GH), dat vooral wordt aangemaakt in de hypofyse, en door insulin-like growth factor I (IGF-I). Bij honden met de ziekte van Cushing (hyperadrenocorticicsme) wordt de hypofyse geremd door de cortisol overmaat, waardoor de aanmaak van GH en daardoor de aanmaak van IGF-I wordt verminderd. Bij honden wordt de ziekte van Cushing in 85% van de gevallen veroorzaakt door een hypofysetumor (en dus in 15% van de gevallen door een bijniertumor). Behandeling bestaat uit het verwijderen van de bijniertumor, medicatie die de bijnier geleidelijk afbreekt, medicatie die de afgifte van cortisol afremt, of verwijdering van de hypofyse (hypofysectomie). Bij hypofysectomie wordt de hypofyse weggenomen en wordt dus ook een belangrijke bron van GH weggenomen. Bij mensen resulteren zowel cortisolovermaat als hypofysectomie Sm in osteoporose (verminderde botmineralisatie) ten gevolge van verminderde 1,25DHCC plasma concentraties met verminderde Ca absorptie en een verhoogde activiteit van paraathormoon (PTH) met verhoogde botresorptie. Hoge PTH activiteit bij honden zorgt voor een gevoel van misselijkheid en verminderde kwaliteit van leven, en was daarom de belangrijkste grondslag voor het onderzoek beschreven in Hoofdstuk 10. Bij volwassen honden met de ziekte van Cushing waren de gehaltes van 1,25DHCC en PTH onveranderd voor en na hypofysectomie vergeleken met gezonde controle honden (Hoofdstuk 10). Dit is mogelijk te verklaren door een veel lagere bot-omzetting bij volwassen honden vergeleken met volwassen mensen. Concluderend hebben honden met de ziekte van Cushing voor noch na hypofysectomie een verhoogde vitamine D behoefte.

Vitamine A is het andere vitamine dat belangrijk is voor het skelet. Zoals is genoemd in Hoofdstuk 6 stimuleert vitamine A de differentiatie van cellen, activeert botoplossende cellen (osteoclasten) en stimuleert het enzym 24-hydroxylase. 24-hydroxylase is nodig voor de vorming van 24,25-dihydroxycholecalciferol (24,25DHCC) en voor de afbraak van

228 Samenvatting

1,25DHCC naar 1,24,25-trihydroxycholecalciferol. Een belangrijke functie van 24,25DHCC is het stimuleren van botaanmaak, zonder bijbehorende botafbraak. De invloed van vitamine A overmaat bij katten is onderzocht in de jaren 60 en 70 van de vorige eeuw, waarbij vooral botnieuwvormingen werden gevonden, meestal periarticulair en aan de halswervels. Alle aanbevelingen voor de hoeveelheid vitamine A voor de productie van kattenvoeding zijn nog steeds gebaseerd op deze oude studies. Subtiele botveranderingen en leverfibrose werden gevonden bij 24 katten die gedurende 18 maanden werden gevoerd met de maximale hoeveelheid vitamine A die op dit moment is voorgeschreven voor kattenvoer (Hoofdstuk 12). Om de klinische relevantie van de botveranderingen, zichtbaar op de röntgenfoto’s, aan te tonen is de toepasbaarheid van kracht plaat analyse onderzocht bij katten (Hoofdstuk 11). Krachtenpatronen van katten verschillen in sommige aspecten duidelijk ten opzichte van honden. De kracht plaat analyse heeft bij katten een hoge reproduceerbaarheid en gevoeligheid waardoor het een objectieve methode is om kreupelheid bij katten te vervolgen. De subtiele botveranderingen bij de katten die veel vitamine A hadden gegeten resulteerde (nog) niet in meetbare kreupelheid op de krachtplaat (Hoofdstuk 12). De klinische relevantie van de leverfibrose die is aangetoond met behulp van immunohistochemische kleuringen is twijfelachtig, aangezien geen van de katten (al) symptomen van leverfalen lieten zien. We kunnen echter niet uitsluiten dat verhoogde opname van vitamine A inderdaad is te relateren aan een voorstadium van leverziekte. Het is opmerkelijk dat matige overmaat van vitamine D (zoals dit voorkomt in natuurlijk voedsel zoals rauwe lever) de nadelige invloeden van vitamine A overmaat op de lever kunnen verminderen, zonder effect te hebben op de botwoekeringen, mogelijk door het anti-fibrotische effect van matige vitamine D overmaat. De bevindingen uit Hoofdstuk 12 zouden een aanzet moeten geven tot een heroverweging van de richtlijnen voor de hoeveelheid vitamine A in kattenvoer, mede in relatie tot de hoeveelheid vitamine D in dit kattenvoer.

Uit de studies beschreven in dit proefschrift kan worden geconcludeerd dat overgewicht, veroorzaakt door een onbalans in energie opname en energie verbruik, schadelijk is Sm voor de gewrichten en dat gewichtsverlies en verhoogde opname van EPA en DHA kan worden toegepast in de behandeling van OA. Extra vitamine D toevoegen aan PMRs is mogelijk niet effectief. Bovendien leidt extra vitamine D niet tot een vermindering van het optreden van LPC bij jonge honden. Ook is het geven van extra vitamine D aan honden met de ziekte van Cushing (hyperadrenocorticisme) niet noodzakelijk. De toevoeging van een beperkte hoeveelheid vitamine D aan katten die veel vitamine A opnemen kan een beschermend effect hebben op de schadelijke effecten van vitamine A op de lever (leverfibrose), zonder nadelige effecten op het skelet.

229 230 Curriculum vitae and list of publications

231 De auteur van dit proefschrift is geboren op 28 april 1979 te Alkmaar. Na het voltooien van het atheneum op het Han Fortmann college in Heerhugowaard is hij in september 1997 gestart met de opleiding Diergeneeskunde in Utrecht, die in december 2003 succesvol is afgerond. Na enkele waarnemingen in eerstelijns gezelschapsdierenartsenpraktijken is hij in dienst gekomen bij Dierenkliniek ’t Ossehoofd te Heerhugowaard, waar hij twee jaar heeft gewerkt als dierenarts gezelschapsdieren. In december 2006 heeft hij een eigen praktijk geopend: Dierenartsenpraktijk Schonauwen te Houten. In september 2008 is hij op het Departement Geneeskunde van Gezelschapsdieren van de Faculteit Diergeneeskunde in Utrecht gestart met de opleiding tot specialist Veterinaire Diervoeding (opleider Prof. Dr. H.A.W. Hazewinkel). Naast de opleiding in Utrecht, zijn stages gelopen bij Dr. A.J. Fascetti (UC Davis, California, USA), Dr. J.J. Wakshlag (Cornell, New York, USA), Dr. J. Kamphues (Hannover, D) en Dr. A.J. German (Liverpool, UK). De specialisatie-opleiding is in september 2011 succesvol afgerond met het ECVCN-examen in Zaragoza waarmee de registratie als Diplomate van the European College of Veterinary and Comparative Nutrition een feit werd. Tijdens zijn specialisatie werd hij in staat gesteld de discipline Klinische Voeding verder vorm te geven en tevens een promotietraject te starten bij de onderzoeksgroep Advances in Veterinary Medicine (coördinator: Prof. Dr. J.W. Hesselink) van het Departement Geneeskunde van Gezelschapsdieren van de Faculteit Diergeneeskunde in Utrecht (begeleiders: Prof. Dr. H.A.W. Hazewinkel, Dr. M.A. Tryfonidou en Dr. A.B. Vaandrager). In maart 2013 heeft hij de basis kwalificatie onderwijs behaald. Hij is getrouwd met Lucas Maria van der Wurff in september 2005. Op dit moment is hij werkzaam als Universitair Docent bij de Faculteit Diergeneeskunde van de Universiteit Utrecht.

The author of this thesis was born on the 28th of April 1979 in Alkmaar. After finishing athenaeum at the Han Fortmann college in Heerhugowaard he started to study Veterinary Medicine at Utrecht University in September 1997. He graduated in December 2003. After several temporal positions in first line companion animal practice he was employed by Dierenkliniek ‘t Ossehoofd in Heerhugowaard, where he worked for 2 years as a veterinarian for companion animals. In December 2006 he opened his private practice; Dierenartsenpraktijk Schonauwen in Houten. In September 2008 he started his residency CV in Veterinary Nutrition (Diplomate of the European College of Veterinary and Comparative Nutrition) (supervisor Prof. Dr. H.A.W. Hazewinkel). Apart from the training programme in Utrecht, his training included externships with Dr. A.J. Fascetti (UC Davis, California, USA), Dr. J.J. Wakshlag (Cornell, New York, USA), Dr. J. Kamphues (Hannover, D) and Dr. A.J. German (Liverpool, UK). He passed the board certifying exam in Zaragoza in September 2011. During his residency he got the opportunity to implement Clinical Nutrition at the Department of Clinical Sciences of Companion Animals, Faculty of Veterinary Medicine, Utrecht University and started his thesis within the research group Advances in Veterinary Medicine (chair: Prof. Dr. J.W. Hesselink, supervisors: Prof. Dr. H.A.W. Hazewinkel, Dr. M.A. Tryfonidou and Dr. A.B. Vaandrager). In March 2013 he was awarded the University Teaching Qualification. He was married to Lucas Maria van der Wurff in September 2005. At the moment he is employed as a University Teacher at the Faculty of Veterinary Medicine, Utrecht University.

232 Curriculum vitae and list of publications

• Bokken, G.C.A.M., Corbee, R.J., Knapen, F. van, Bergwerff, A.A., 2003. Immunochemical detection of Salmonella group B, D and E using an optical surface plasmon resonance biosensor. FEMS Microbiology Letters 222, 75-82. • Corbee, R.J., Tryfonidou, M.A., Beckers, I.P., Hazewinkel, H.A.W., 2012. Comparison of the composition and the use of puppy milk replacers in German Shepherd puppies in the Netherlands. Journal of Animal Physiology and Animal Nutrition 96, 395-402. • Corbee, R.J., Booij-Vrieling, H.E., Lest, C.H.A. van de, Penning, L.C., Tryfonidou, M.A., Riemers, F., Hazewinkel, H.A.W., 2012. Inflammation and wound healing in cats with chronic gingivitis stomatitis after extraction of all premolars and molars were not affected by feeding of two diets with different omega-6:omega-3 polyunsaturated fatty acids ratios. Journal of Animal Physiology and Animal Nutrition 96, 679-688. • Corbee, R.J., Tryfonidou, M.A., Meij, B.P., Kooistra, H.S., Hazewinkel, H.A.W., 2012. Vitamin D status before and after hypophysectomy in dogs with pituitary-dependent hypercortisolism. Domestic Animal Endocrinology 42, 43-49. • Corbee, R.J., Kerkhoven, W., van, 2012. Nutrition of the hospitalized patient (Article in Dutch). Tijdschrift voor Diergeneeskunde 157, 384-390. • Corbee, R.J., Tryfonidou, M.A., Barnier, M.M.C., Lest, C.H.A. van de, Hazewinkel, H.A.W., 2013. The effect of dietary long-chain omega-3 fatty acids supplementation on owner’s perception of behavior and locomotion in cats with naturally occurring osteoarthritis. Journal of Animal Physiology and Animal Nutrition 97, 846-853. • Corbee, R.J., 2013. Obesity in show dogs. Journal of Animal Physiology and Animal Nutrition 97, 904-910. • Corbee, R.J., 2014 Obesity in show cats. Journal of Animal Physiology and Animal Nutrition, accepted. DOI: 10.1111/jpn12176. • Corbee, R.J., Kerkhoven, W.J.S., van, 2014. Nutritional support of dogs and cats after surgery or illness. Open Journal of Veterinary Medicine 4, 44-57. • Corbee, R.J., Maas, H., Doornenbal, A., Hazewinkel, H.A.W., 2014. Ground reaction forces of walking cats; assessment and comparison with walking dogs. The Veterinary Journal, accepted. • Corbee, R.J., Tryfonidou, M.A., Grinwis, G.C.M, Wolschrijn, C.F., Lau, S.F., Gorissen, B.M.C., Vaandrager, A.B., Hazewinkel, H.A.W., 2014. Dietary vitamin D supplementation during CV early growth does not protect against medial coronoid disease in Labradors. The Veterinary Journal, submitted. • Corbee, R.J., Tryfonidou, M.A., Grinwis, G.C.M., Schotanus, B.A., Molenaar, M.R., Voorhout, G., Vaandrager, A.B., Hazewinkel, H.A.W., 2014. Chronic vitamin A supplementation in cats results in minor skeletal changes and liver fibrosis; the latter was counteracted by moderate vitamin D supplementation. The Veterinary Journal, submitted. • Corbee, R.J., Vaandrager, A.B., Kik, M.J.L., Molenaar, M.R., Hazewinkel, H.A.W., 2014. Cutaneous vitamin D synthesis in carnivorous species. Journal of Animal Physiology and Animal Nutrition, submitted.

233 234 Acknowledgements

235 Allereerst wil ik mijn promotor Professor Herman Hazewinkel bedanken. Tijdens mijn eindgesprek na de co-schappen in 2003 heeft u aangegeven bezig te zijn met het opstarten van een specialisatie Klinische Voeding. Ik was uiteraard erg enthousiast en hoewel het enige tijd heeft geduurd, ben ik daar uiteindelijk in 2008 mee begonnen. Uw uitgebreide netwerk heeft mij in contact gebracht met vele inspirerende mensen die mij de fijne kneepjes van het vak hebben geleerd. Het examen in Zaragoza in 2011 werd succesvol afgerond, waarna ik aan mijn promotietraject kon gaan beginnen. Naast alle andere taken was het voor mij soms lastig om de focus op het proefschrift te houden; u hield mij als geen ander op de weg, en zie hier het resultaat: een proefschrift met voeding en orthopedie (uw beide specialisaties) als onderwerp.

Daarnaast wil ik mijn tweede promotor Professor Jan-Willem Hesselink bedanken. U heeft de multidisciplinaire aanpak “over grenzen van mogelijkheden” geïntroduceerd in de kliniek. Omdat wij beiden zijn gespecialiseerd in een diersoortoverschrijdende discipline delen wij de interesse om naar het grote geheel te kijken en daarbij nieuwe mogelijkheden te ontdekken. Aangezien de voeding raakvlakken heeft met zeer veel andere disciplines zie ik er naar uit om nog lang met u samen te mogen werken.

Onmisbaar waren ook mijn co-promotoren. Doctor Marianna Tryfonidou, Het was fantastisch om met u samen te mogen werken (σας ευχαριστώ). De frisse blik en de snelle correcties die de tekst een stuk gemakkelijker leesbaar maakten, hebben een zeer positieve bijdrage geleverd aan het eindresultaat. Doctor Bas Vaandrager, uw biochemische kennis was van grote waarde bij de totstandkoming van dit proefschrift. De vele vitamine A en vetzuur analyses heb ik met u en de andere medewerkers van het Departement Biochemie en Celbiologie mogen uitvoeren, waarvoor hartelijk dank.

Uiteraard mogen ook mijn vele co-auteurs in dit dankwoord niet ontbreken. Zoals u kunt zien lopen de onderwerpen in dit proefschrift nogal uiteen en op alle verschillende gebieden werden experts om hun medewerking gevraagd. Zoals Doctor Lau al aangaf in haar proefschrift: “Without you nothing is possible, with you nothing is impossible”.

Velen hebben mij geholpen met het verzamelen van monsters, het uitvoeren van de Ak analyses, het meedoen met de onderzoeken. Doctor Henri Heuven en Doctor Erik Teske hebben mij geholpen met de statistiek, waarvoor hartelijk dank.

Daarnaast zijn de dierverzorgers en het ondersteunend personeel belangrijk geweest voor de dagelijkse verzorging van de onderzoeksdieren en de assistentie bij het verzamelen van monsters. In het bijzonder wil ik Inge van Duiven en Harry van Engelen bedanken voor hun inzet. Daarnaast de medewerkers van het Universitair Veterinair Diagnostisch Laboratorium voor het verwerken van alle bloedmonsters, de medewerkers van de apotheek voor het vervaardigen van de oliën, de medewerkers van het onderzoekslaboratorium van Gezelschapsdieren voor de immuunhistochemie (in het bijzonder Ingrid Rombouts-van Gils en Jeannette Wolfswinkel), en de

236 Acknowledgements

medewerkers van de diagnostische beeldvorming voor het maken en analyseren van de röntgenfoto’s, CT-scans en echografiebeelden.

Onderzoek doen kost geld, en ook de opleiding tot specialist is kostbaar. Ik wil alle sponsoren bedanken voor hun bijdrage. In het bijzonder Doctor Vincent Biourge (Royal Canin Europe), Dhr. Dennis Cordes en Drs. Margriet Bos (Royal Canin Nederland); zonder jullie steun had ik geen voedingsspecialist kunnen worden (merci beaucoup). Doctor Luc Janssens, Mvr. Irene Rol en Mvr. Marlies Blom (Doils, NML Health), hartelijk dank voor het beschikbaar stellen van de supplementen en de placebo’s.

De organisatie van de Winner tentoonstelling en de organisatie van de Mundikat kattenshows: hartelijk dank dat u mij mijn onderzoek hebt laten uitvoeren op de door u georganiseerde shows. Ik dank de Raad van Beheer op Kynologisch gebied in Nederland voor het geven van een passend vervolg aan de door mij gegenereerde resultaten.

Tijdens mijn ontwikkeling zijn een aantal mensen onmisbaar geweest. Ik wil mijn proefschrift opdragen aan mijn ouders. Al op jonge leeftijd maakten jullie mij ervan bewust dat je altijd het maximale uit jezelf moet halen. Jullie geloofden altijd in de goede afloop en waren mijn steun en toeverlaat. Een andere belangrijke steun en toeverlaat is mijn echtgenoot Lucas van der Wurff. Dit proefschrift is mede door jou tot stand gekomen; jij mocht altijd als eerste de teleurstellingen opvangen en de successen meevieren. Daarnaast mijn familie en vrienden. In het bijzonder de “heren van de ronde tafel”; mijn vaste vriendenclub sinds het begin van de studie in 1997. Ook mijn collega’s, en in het bijzonder mijn kamergenoten: bedankt voor de mooie momenten en de interessante gesprekken. Niet voor niets zijn mijn paranimfen René Steinmeijer (een van de “heren van de ronde tafel”) en Erik Wouters (mijn kamergenoot gedurende een aantal jaren).

Het Engels werd gecorrigeerd door Doctor Jane Sykes en door Doctor Joseph Wakshlag. Dear Jane and dear Joe, thank you very much for your contributions to my PhD-thesis.

237 Ak

238