Leukoencephalomyelopathy in cats linked to abnormal fatty acid composition of the white matter of the spinal cord and of irradiated dry cat food

Abstract Four outbreaks of leukoencephalomyelopathy in colonies of SPF cats on a long‐term diet of irradiated dry cat food were observed in the Netherlands between 1989 and 2001. As a primary defect in myelin formation was suspected to be the cause of the disease and myelin consists mainly of lipids and their fatty acids, we investigated the fatty acid composition of the white matter of the spinal cord of affected and control cats and of irradiated and non‐irradiated food. The irradiated food had low levels of alpha‐linolenic acid compared to linoleic acid as well as a high total omega‐6:omega‐3 ratio of 7:1 in the irradiated and of 2:1 in the non‐irradiated food. The white matter of the spinal cord showed low levels of linoleic acid and absence of alpha‐linolenic acid in affected cats as well as absence of lignoceric and nervonic acid in both affected and control cats. These abnormalities in fatty acid composition of the white matter of the spinal cord may reflect an increased need for alpha‐linolenic acid as a substrate for longer chain omega‐3 fatty acids to compose myelin and thus indicate a particular species sensitivity to dietary deficiency in omega‐3 polyunsaturated fatty acids, particularly alpha‐linolenic acid in cats. Our findings indicate that abnormalities in fatty acid metabolism in myelin play an essential role in the pathogenesis of this acquired form of leukoencephalomyelopathy in cats.


| INTRODUC TI ON
Leukoencephalomyelopathy has first been described in young cats suffering from a progressive form of ataxia between 1969 and 1980.
These cats originated from specific pathogen-free (SPF) colonies and the neurological condition affected up to 40% of litters in a colony (Palmer & Cavanagh, 1995). Although initially an infectious or toxic cause was suggested, extensive investigations for such aetiologies were consistently negative. In a series of outbreaks of a clinically and pathologically similar syndrome between 1998 and 2001, reported from Ireland, 25%-40% of cats in SPF colonies were affected (Cassidy et al., 2007). The authors provided initial circumstantial evidence that the condition was associated with the exclusive feeding of the cats of a gamma-irradiated dry cat food. Their hypothesis was further substantiated as after supplementation and replacement of the irradiated food with pasteurized proprietary tinned cat food no further cases occurred, and definitely proven by the experimental induction of leukoencephalomyelopathy in cats by long-term (140-174 days) feeding of a gamma-irradiated dry cat food (Caulfield et al., 2009). Similar cases of leukoencephalomyelopathy, both clinically and pathomorphologically, have been observed in Australia between 2008 and in 87 private owned cats eating imported, irradiated dry cat food (Child, Foster, Fougere, Milan, & Rozmanic, 2009). From the studies of Duncan, Brower, Kondo, Curlee, and Schultz (2009) on demyelinization and remyelinization of the central nervous system in cats on a long-term diet of irradiated dry cat food, it is evident that the myelin sheath is primarily affected and axonal damage is initially absent.
Although the irradiated dry cat foods have decreased vitamin A levels and increased peroxide levels, and increased peroxide levels were found in the liver and spinal cords of the affected cats (Caulfield et al., 2009), a definite cause/effect relationship could not be established.
Depletion of vitamin A after gamma-irradiation has been suggested as a possible contributing factor of leukoencephalomyelopathy in cats (Cassidy et al., 2007;Caulfield et al., 2009) However, the reported (6,230-6,900 IU/kg dry matter basis) vitamin A concentration in the irradiated food in these cases was low, but within normal reference range for the adult cat according to FEDIAF nutritional guidelines publication December 2018 (3,330-400,000 IU/kg dry matter basis). Moreover, cats with leukoencephalomyelopathy associated with irradiated food fail to develop clinical signs or lesions that are typical for vitamin A deficiency, such as obvious squamous metaplasia of the epithelia of conjunctiva, trachea, salivary ducts and uterus (Gershoff, Andrus, Hegsted, & Lentini, 1957).
In a more recent study, the peroxide levels in commercial dry animal diets were increased 14-to 25-fold after high single-dose gamma-irradiation of 38.4-48.7 kGy, and cats with leukoencephalomyelopathy had increased peroxide levels in liver and spinal cords (Caulfield, Cassidy, & Kelly, 2008). Hydrogen peroxide is toxic to many cells if it comes into contact with reduced forms of certain metal ions, forming the highly reactive hydroxyl radical that acts to cause lipid peroxide formation (Halliwell & Gutteridge, 1985). However, the mechanism by which changes induced in foods then result in damage to the white matter of the spinal cord and brain in cats is not clear (Child et al., 2009). Moreover, irradiated dry dog and rodent diets which have similar levels of peroxides never caused nervous signs or lesions in these animals (Caulfield et al., 2008).
The present paper describes the pathomorphological findings in four outbreaks of leukoencephalomyelopathy in colonies of SPF cats fed irradiated dry cat food in the Netherlands between 1989 and 2001. Because a definite cause for the disease has not been established until now, and the demyelinating lesions and their distribution throughout the white matter of the central nervous system in these cats show resemblance to primary demyelination due to defective or improper myelin formation as in Canavan's disease and adrenomyelopathy in man and associated experimental mouse models (Baumann & Pham-Dinii, 2001); polyunsaturated fatty acids, including essential fatty acids (EFA), are highly vulnerable to irradiation due to an increase in lipid peroxide formation (Hammer & Wills, 1979), reflected by higher thiobarbituric acid reactive substances (TBARS) and nuclear magnetic resonance (NMR) values after irradiation (De Oliveira Silva, Mársico, Oliveira de Jesus, Guimarães, & Sloboda Cortez, 2011) mostly affecting polyunsaturated fatty acids resulting in more saturated fatty acids (Javan & Motallebi, 2015), and increased formation of trans fatty acids (Yilmaz & Geçgel, 2007); nervous tissue concentrates omega-3 polyunsaturated fatty acids in myelin (Lenox, 2015); lipid composition and fatty acid profiles of myelin of rat brain vary in response to the consumption of different fats (Bourre et al., 1984;Srinivasarao, Vajreswari, Rupalatha, & Narayanareddy, 1997); and induction of EFA deficiency in rats resulted in both a decrease of EFA in myelin but also vacuolation/myelin splitting in the optic nerve (Trapp, 1977), we hypothesized that abnormalities in fatty acid metabolism in myelin might play an essential role in the pathogenesis of this acquired form of leukoencephalomyelopathy in cats. Therefore, we investigated in addition the fatty acid profile of the white matter of the spinal cord of affected cats and control animals, and of irradiated and non-irradiated food.

| Animals
In July-August 1989, 12 out of 33 cats in a breeding colony developed progressive paresis/ paralysis and ataxia. The colony was maintained unvaccinated in a maximum barrier production unit and received irradiated dry cat food. Cats were between 2 months and 2 years of age, and affected animals were not genetically related and included both males and females. Cats otherwise had normal appetite, water consumption, reproduction and behaviour. From July to October 1993, a second outbreak affecting six out of 21 cats occurred in a group of SPF cats from the same breeding facility as well as in six animals transferred to a research laboratory. Also, these cats were maintained unvaccinated in a maximum barrier production unit and received irradiated dry cat food. In June-July 1997, a third outbreak was observed in the same breeding colony and affected 40 out of 60 females and males. A fourth outbreak with similar symptoms was observed in September-October 2000, in a different breeding colony consisting of 20 females and three males, and finally affecting 80% of the animals, as well as in January 2001 in a group of eight cats transferred from this colony to a research laboratory for a behavioural study. Also, these animals were held under SPF conditions and received irradiated dry cat food. In all these outbreaks severely affected animals died or were euthanized.
Mildly affected animals showed gradual improvement/disappearance of symptoms within 4-10 months after receiving a different batch of food. The three control cats, which were used for fatty acid determinations of the spinal cord white matter, were a private owned queen of 10 months old with acute cardiac decompensation after ovariohysterectomy and two healthy adult queens used as control cats in an experimental feline infectious peritonitis model study.

| Post-mortem examination
Thirteen severely affected cats from the various outbreaks between 1989 and 1997, as well as six out of eight cats from the outbreak in January 2001, the latter all euthanatized after clinical signs developed within the group, were available for post-mortem examination. Complete post-mortem examination was performed by board-certified veterinary pathologists. Samples for histological examination were taken from tonsil, mandibular or pre-scapular lymph node, (para)thyroid gland, thymus, heart, lung, liver, spleen, kidney, adrenal gland, stomach (fundus), pancreas, duodenum, colon, ovary and uterus respectively testicle and skeletal muscle (M. semitendinosus) and in the cats from the 2001 outbreak also from subcutaneous (inguinal) and perirenal fat tissue.
For neuropathological examination, brain, spinal cord including spinal ganglia, right brachial and sciatic nerves, coeliac ganglion, trigeminal ganglion and pituitary gland were fixed by immersion in 10% neutral, phosphate-buffered, formalin. After routine dehydration and embedding in paraffin, 4 µm sections were made and stained with haematoxylin and eosin (HE). Specific staining methods used in selected slides from the nervous system were toluidine blue and immune-histochemical stains for glial fibrillary acidic protein (GFAP), ionized calcium binding adaptor molecule 1 (IBA-1; a microglial and histiocyte marker) and alpha-smooth muscle actin (αSMA) using standard procedures.

| Fatty acid determinations in spinal white matter and food
Spinal white matter was collected from three affected cats from the 1997 outbreak and six affected cats from the outbreak in 2001 as well as from the three control cats. Spinal cord material was frozen in liquid nitrogen and stored at −70°C until use. Immediately prior to analysis the material was thawed and ground. Also, a sample of the irradiated food associated with the 2000 outbreak (SSNIFF, 2.5Mrad = 25 kGy) and of a standard dry cat food from the same manufacturer were obtained and ground prior to analysis.
Cholesteryl esters (CE) 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 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 × 3 mm; Phenomenex, CA, USA). Internal standards were used for comparison.

| Pathomorphological findings
All animals were in good bodily condition, and no gross abnormalities were seen.
Histologically, lesions were consistently present in the central nervous system and comparable in distribution and severity in all severely affected cats. In both the brain and the spinal cord, lesions were restricted to the white matter, and consisted of bilateral symmetric vacuolation, loss of myelin and variously associated astrocytosis, microgliosis and capillary hyperplasia. The vacuoles often contained intact and sometimes swollen axons but also loss of axons associated with cellular debris and/or myelophages was observed

| D ISCUSS I ON
The irradiated food-induced leukoencephalomyelopathy in cats as demonstrated in our study has some special characteristics and thereby differs from many other forms of demyelination. It is an acquired disease that affects only cats and develops after feeding the irradiated food for four to six months. The disease, at least in the early clinical stage, is morphologically characterized by myelin vacuolation and demyelination with some myelophages but in general in the presence of intact axons, as well as by remyelinization with thin myelin sheaths around intact axons. When the cats at this stage are returned to a normal food, the animals show complete clinical recovery after 6-8 weeks although after 12 and 24 months these recovered cats still show thin myelin sheaths characteristic for remyelinated axons (Duncan, Marik, Broman, & Heidari, 2017). As already suggested by Duncan et al. (2009), the disease therefore may be regarded as a primary disease of the myelin sheath/oligodendrocyte. Defective myelination may disrupt axo-glial cell interactions, axolemmal organization and electrical conduction and connectivity and hence in chronic cases may cause axonal damage and loss affecting the specific biophysical properties of myelin as an electrical insulator (Piaton, Gould, & Lubetzki, 2010). The observed reactive astrocytosis and microglial and endothelial proliferation most likely are secondary phenomena.
The myelin sheath is an extended highly specialized plasma membrane synthesized in the central nervous system by oligodendrocytes. One of the prominent biochemical characteristics that distinguishe myelin from other membranes is its high lipid-to-protein ratio. Lipids account for at least 70% of the dry weight of myelin membranes. Much of the structure and function of myelin is dependent on its lipid content. The lipid characteristics of myelin together with the presence of myelin-specific proteins are likely required for myelin wrapping and compaction to confer to the specific biophysical properties of myelin as an electrical insulator (Chrast, Saher, Nave, & Verheyen, 2011). Mature compacted myelin in rats is markedly enriched in saturated and monosaturated fatty acids, especially lignoceric acid (C24:0) and nervonic acid (C24:1n-9), whereby the maturity and compaction of myelin is assessed by the ratio of lignoceric acid  (C24:0) + nervonic acid (C24:1n-9) to stearic acid (C18:0) + oleic acid (C18:1n-9). In fact, the accumulation of these marker very long-chain fatty acids (VLCFA) determines the process of myelin maturation in rats (Chrast et al., 2011;Srinivasarao et al., 1997).
A remarkable finding with respect to the fatty acid composition of the white matter of the spinal cord in affected animals compared to normal cats were the low levels of linoleic acid (C18:2n-6) and even absence of alpha-linolenic acid (C18:3n-3) and the high levels of docosadienoic acid (C22:2n-6), docosatetraenoic acid (C22:4n-6) and unknown fatty acids in the affected cats. As myelin composition is directly influenced by the composition of dietary fats (Bourre et al., 1984;Srinivasarao et al., 1997), cats, it is regarded as such for growing kittens (FEDIAF nutritional guidelines publication December 2018). The total omega-6:omega-3 ratio in the irradiated food is 7:1, whereas nutritionists typically recommend ratios between 2:1 and 1:1 for cat food. Such increased total omega-6:omega-3 ratio in the irradiated food might contribute to a relative deficiency of alpha-linolenic acid (C18:3n-3), as there is competition for metabolic enzymes with omega-6 fatty acids such as linoleic acid (C18:2n-6). Different omega-6:omega-3 ratios are reported for different indications (Simopoulos, 2002), and not only omega-6:omega-3 ratio is important, as there needs to be a certain amount of eicosapentaenoic acid (C20:5n-3) and docosahexaenoic acid (C22:6n-3) present in the diet to achieve effective plasma levels (Hall, Picton, Skinner, Jewell, & Wander, 2006).
Oligodendrocytes in the adult are thought to primarily rely on uptake to maintain their fatty acids pool (Dimas, 2017).
Oligodendrocytes with a shortage in lipid synthesis appear to be supported by horizontal lipid flux from astrocytes (Camargo et al., 2017). Failure of astrocytes to provide this support results in increased uptake of circulating usually dietary lipids by myelinating oligodendrocytes (Camargo et al., 2017). In case of a deficient dietary supply of specific lipid components like n-3 polyunsaturated fatty acids, these may be released by lipolysis from perineural adipocytes TA B L E 2 Fatty acid composition of the irradiated (SSNIFF 2,5Mrad = 25 kGy) and non-irradiated food (from same manufacturer) expressed as percentage of total fatty acids  Increased peroxide levels in the spinal cord of affected cats as demonstrated by Caulfield et al. (2008) in our opinion are not a cause but a secondary phenomenon as abnormal fatty acid composition in murine oligodendrocytes has been described to induce oxidative stress and overproduction of reactive oxygen species (Baarine et al., 2012).
However, the myelin sheath has been regarded for a long time as an inert insulating structure, it has now become clear that myelin is metabolically active. Even if the turnover of lipids in myelin is slow with a half-life of several weeks to months, there is a constant exchange of molecules (Schmitt, Castelvetri, & Simons, 2015;Simons & Nave, 2015). This may explain the long, 4-6 months, period for the development of clinical disease in cats exposed to the irradiated dry cat food.
Apart from domestic cats also other wild felidae may be more sensitive to these alterations in dietary fatty acid pattern as a comparable form of leukoencephalomyelopathy has been described in captive cheetahs and other large felids (Brower et al., 2014;Palmer et al., 2001). Unfortunately, in these cases the diet composition was unknown. However, nutritional deficiencies are common in wild animals that are kept in captivity (Crissey, 2005), so a deficient or unbalanced fatty acid pattern of the diet also might be at the origin of the leukoencephalomyelopathy in these wild felids.

ACK N OWLED G EM ENTS
We wish to thank Tim van Olmen, Monique van Wolferen, Henny Karelse and Louis van den Boom for technical support.