Arachidonic acid (AA) is a PUFA present in the phospholipids of cell membranes, and it is particularly abundant in the retina and brain(Reference Arterburn, Hall and Oken1, Reference Semba and Semba2). AA in the human body comes from dietary sources such as egg yolk(Reference Semba and Semba2), or it is synthesised from linoleic acid(Reference Le, Meisel and de Meijer3). Neurological health requires sufficient levels of DHA and AA(Reference Davis-Bruno and Tassinari4). Early infancy may be a critical time when visual development and brain development are susceptible to the effects of inadequate stores or a deficient intake of essential fatty acids, DHA and AA(Reference Saste, Carver and Stockard5). AA drives postnatal neurogenesis and elicits a beneficial effect on prepulse inhibition in Pax6 knockout mice characterised by impaired postnatal neurogenesis(Reference Maekawa, Takashima and Matsumata6). Several randomised clinical trials of supplemental DHA and AA have been conducted in full-term infants, and infants who received the supplementation had better visual-evoked potential acuity when compared with the control groups(Reference Semba and Semba2, Reference Uauy, Hoffman and Peirano7). Recently, nutrient-based preventive treatments for development and progression on human retinal diseases have been focused, and large-scale clinical studies, such as the Age-Related Eye Disease Study, have been conducted(8).
Inherited night blindness is a fairly widespread disease in humans, and it affects approximately one in 5000 individuals worldwide. A common form of inherited blindness is retinitis pigmentosa (RP), which is a degenerative pigmentary retinopathy that is non-inflammatory, bilateral and progressive(Reference Hartong, Berson and Dryja9, Reference Shintani, Shechtman and Gurwood10). RP is a heterogeneous group of inherited retinal disorders caused by more than 160 different mutations of genes encoding proteins with remarkably diverse functions that are known to cause photoreceptor degeneration(11). In many RP cases, clinical symptoms typically start in the early teenage years and severe visual impairment occurs by ages 40–50 years(Reference Ferrari, Di Iorio and Barbaro12). Patients with Usher syndrome, roughly one-sixth of patients with RP, begin to lose their vision in their first decade of life(Reference Williams13). This type can be diagnosed before birth by using genetic methods and/or a pedigree diagram. In these cases, the initiation of therapy before birth may prevent or delay retinal damage in young people with RP.
Animal models of retinal degeneration are important for elucidating the mechanisms of human RP(Reference Rivas and Vecino14, Reference Sancho-Pelluz, Arango-Gonzalez and Kustermann15) and exploring potential treatments(Reference Shintani, Shechtman and Gurwood10, Reference Tsubura, Yoshizawa and Kuwata16, Reference Yoshizawa and Tsubura17). Mutant mice used as models for RP include mice carrying the rodless retina or retinal degeneration (rd) gene, or the retinal degeneration slow (rds) gene; these mice develop photoreceptor cell death caused by apoptosis(Reference Sancho-Pelluz, Arango-Gonzalez and Kustermann15). In addition to inherited RP models, N-methyl-N-nitrosourea (MNU), an alkylating agent that targets photoreceptor cells, rapidly induces retinal damage via apoptosis in several animal species(Reference Tsubura, Yoshizawa and Kuwata16, Reference Yoshizawa and Tsubura17). MNU-induced retinal damage is due to selective 7-methyldeoxyguanosine DNA adduct formation in photoreceptor cell nuclei followed by photoreceptor cell death via an apoptotic mechanism(Reference Yoshizawa, Nambu and Yang18). The MNU-induced apoptosis cascade involves the up-regulation of Bax protein, the down-modulation of Bcl-2 protein and the activation of caspase-3, -6 and -8(Reference Yoshizawa, Nambu and Yang18). Within 7 d after an adequate dose of MNU is administered, active signs of photoreceptor degeneration are distinct due to photoreceptor cell loss, and the inner nuclear layer is either in direct contact with the choroid or is separated from it by a few layers of cells. This model mimics the cell death process in human RP and can be used for studies of therapeutic intervention. The goal of the present study is to elucidate the effect of prenatal and postnatal dietary AA on MNU-induced retinal degeneration in young Lewis rats.
Experimental materials
Animal procedures
The protocol and all animal procedures used in the present experiment were approved by the Animal Care and Use Committee of Kansai Medical University and were in accordance with the guidelines for animal experimentation at the Kansai Medical University. A total of thirty-one male and female SPF/VAF rats (LEW/CrlCrlj) at 10–12 weeks of age were purchased from Charles River Japan. In our previous study, we confirmed that MNU induces retinal degeneration in both male and female rats, without the sex difference in incidence and severity. Lewis inbred rats have a similar sensitivity of retinal damage to the other strain. Therefore, we selected male Lewis rats in the present study. Rats were maintained in specific pathogen-free conditions and received free access to water and CE-2-modified diets containing different doses of AA. The animals were housed in plastic cages with paper-chip bedding (Paper Clean; SLC) in an air-conditioned room at 22 ± 2°C and 60 ± 10 % relative humidity with a 12 h light–dark cycle. Illumination intensity in the cages was less than 60 lux. In each dietary group, each male was mated with six to nine females. Offspring were culled to a maximum of ten per dam, and the dams were maintained on their respective diets during the 21 d lactation period and a post-weaning period of up to 7 d. At the age of 21 d, male pups were selected and used in the present study (Fig. 1). In each dietary group, necropsy was done simultaneously for five vehicle-treated rats and four to eight MNU-treated rats.
Schema of the experimental protocol. AA, arachidonic acid; i.p., intraperitoneal; HE, haematoxylin and eosin; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP digoxigenin nick-end labelling; MNU, N-methyl-N-nitrosourea.
Chemical and dose formulation
MNU was obtained from Sigma-Aldrich and was kept at − 80°C in the dark. The MNU solution was dissolved in physiological saline containing 0·05 % acetic acid just before use. MNU (50 mg/kg) or vehicle (physiological saline containing 0·05 % acetic acid) was administered once by intraperitoneal injection(Reference Yoshizawa, Nambu and Yang18, Reference Yoshizawa, Kuwata and Kawanaka19).
Arachidonic acid-supplemented diet
AA (CABIO, purity of 40·4 % by analysis) was purchased from Cargill Alking Bioengineering (Wuhan) Company Limited and purified by T. M. for use. Diets with three concentrations of AA (0·1, 0·5 and 2·0 % (w/w)) were semi-purified based on the modified CE-2 formulation (CLEA Japan, Inc.). The basal diet consisted of modified CE-2. The different concentrations of AA were added to prepare each experimental diet. GC analyses of the fatty acid composition of the diets are shown in Table 1(Reference Hussein, Fedorova and Moriguchi20). The total fatty acid volumes were 86·75, 131·96, 126·70 and 126·63 μg/mg diet for the basal diet ( < 0·01 % (w/w) AA), 0·1, 0·5 and 2·0 % AA diet, respectively. These diets were exposed to γ-rays (30 Gy), and formulated by CLEA Japan, Inc. The diets were stored at 4°C to prevent lipid oxidation before use.
Fatty acid composition of the diets
AA, arachidonic acid; ND, not determined.
Experimental procedures
Male Lewis rats were exposed to the basal diet or an experimental diet (0·1, 0·5 and 2·0 % AA) from fertilisation to killing. At 21 d of age, these rats received an intraperitoneal injection of vehicle (physiological saline) or 50 mg MNU/kg (Fig. 1). At 24 h (five rats in each dietary group) or 7 d (four to eight rats in each dietary group) after the MNU treatment, rats were anaesthetised with isoflurane (Forane®; Abbot Japan) and killed by exsanguination from aortic transection. During the experiment, all pups were observed daily for clinical signs of toxicity and were weighed at the time of MNU treatment and on the day of killing. Both eyes were quickly removed at the time of killing, and complete necropsies were conducted on all animals. Food consumption and the body weight of the dams were measured once per week to estimate the actual dosage of AA during the mating, pregnancy and lactation periods.
Fatty acid analysis of retina and serum
For the analysis, retinas and serum samples were collected from each of the rats killed 7 d after injection of the vehicle. Retinas and blood samples were stored at − 80°C before lipid extraction and the measurement of fatty acid composition. The fatty acid composition was measured as described previously with minor modifications(Reference Harauma and Moriguchi21, Reference Masood, Stark and Salem22). Tissue samples were thawed, weighed and homogenised in methanol–hexane and methylated in acetyl chloride. Varying amounts of internal standards, methyl docosatrienoate (22 : 3n-3 for retina) and methyl tricosanoic (23 : 0 for serum), were added to each sample to compensate for differences in tissue weight and lipid concentration. As an aid in preventing lipid oxidation during the procedures, butylated hydroxytoluene was added to methanol. The hexane extracts were concentrated into microvials for GC injection. Fatty acid methyl esters were analysed with an Agilent 7890A GC (Agilent Technologies) equipped with a split injector, an Agilent 7693A ALS automatic liquid sampler and an flame ionisation detector. The instrument was controlled and data were collected using Agilent ChemStation (Rev.B.04.01.SP1; Agilent Technologies).
Tissue fixation and processing
From the saline-treated rats that were fed each diet, one eye was fixed overnight in methacarn. From the MNU-treated rats, one eye was fixed overnight in 10 % neutral buffered formalin, and the other eye was fixed overnight in methacarn. Subsequently, the eyes were embedded in paraffin, sectioned at a thickness of 4 μm, and stained with haematoxylin and eosin. Ocular sections were cut along a line parallel to the optic axis and nerve (including the ora serrata).
Morphometric analysis of retinal thickness, photoreceptor cell ratio and retinal damage ratio
Haematoxylin and eosin-stained sections and terminal deoxynucleotidyl transferase-mediated dUTP digoxigenin nick-end labelling (TUNEL)-stained sections of the retina were scanned with a high-resolution digital slide scanner (NanoZoomer 2·0 Digital Pathology; Hamamatsu Photonics) to prepare digital images. The image files were opened in colour mode using NDP.view software (Hamamatsu Photonics). The total retinal thickness (from the internal limiting membrane to the pigment epithelium), inner retinal thickness (from the internal limiting membrane to the outer plexiform layer) and outer retinal thickness (from the outer nuclear layer to the pigment epithelial cell layer) were individually measured from methacarn-fixed haematoxylin and eosin-stained slides by using NDP.view, as described previously(Reference Yoshizawa, Kuwata and Kawanaka19, Reference Yoshizawa, Sasaki and Uehara23). The measurements were conducted at the central retina (approximately 400 μm from the optic nerve) and peripheral retina (approximately 400 μm from both sides of the ciliary bodies). To further evaluate the photoreceptor cell loss, the photoreceptor ratio was calculated as:
$$\begin{eqnarray} Photoreceptor\ ratio = (outer\ retinal\ thickness/total\ retinal\ thickness)\times 100. \end{eqnarray}$$
To determine the area of retinal damage, the entire length of the retina and the length of the damaged area in haematoxylin and eosin preparations were measured. Damaged retina was designated as the presence of less than four rows of photoreceptor nuclei in the outer nuclear layer(Reference Yoshizawa, Kuwata and Kawanaka19, Reference Yoshizawa, Sasaki and Uehara23), and the retinal damage ratio was calculated as:
$$\begin{eqnarray} Retinal\ damage\ ratio = (length\ of\ damaged\ retina/whole\ retinal\ length)\times 100. \end{eqnarray}$$
The apoptotic index was calculated from TUNEL-stained slides by determining the number of apoptotic nuclei per microscopic field at three central locations and three peripheral locations of the retina. The images were captured with a 40-fold objective magnification. Histopathological and morphometrical evaluations were performed by a toxicological pathologist certified by the Japanese Society of Toxicologic Pathology and the International Academy of Toxicologic Pathology (K. Y.) and an ophthalmologist certified by the Japanese Ophthalmological Society (M. K.), according to the previously defined histopathological terminology and diagnostic criteria(Reference Yoshizawa, Nambu and Yang18, Reference Yoshizawa, Kuwata and Kawanaka19).
Terminal deoxynucleotidyl transferase-mediated dUTP digoxigenin nick-end labelling staining
Formalin-fixed retinal sections obtained from the rats (basal, 0·5 and 2·0 % AA) that were killed 24 h after the MNU treatment were used for cell death analysis. Cell death was evaluated by TUNEL using an in situ apoptosis detection kit (ApopTag®; Millipore)(Reference Yoshizawa, Kuwata and Kawanaka19, Reference Yoshizawa, Sasaki and Uehara23).
Statistical analysis
All discrete values, expressed as means with their standard errors, were analysed using the two-tailed independent Student's t test for unpaired samples after confirming the homogeneity of variances (Excel 2007; Microsoft). The results presented include comparisons between the basal diet-fed rats and rats fed the different doses of the AA-supplemented diet in both MNU-treated and vehicle-treated groups. P values < 0·05 were considered to show statistical significance.
Results
General remarks
No deaths occurred, and no clinical signs or symptoms were evident in any of the pups or dams during the experimental period. The AA diets did not influence body-weight gain (the growth rate) in pups or cause weight changes in dams with or without the MNU treatment; however, the growth rate in MNU-treated pups tended to be lower than that in vehicle-treated pups (data not shown).
Estimated intake of arachidonic acid
In the pregnancy and lactation periods, the AA intake of dams was 4·7 and 9·4 mg/kg per d in the basal diet group; 77·7 and 242·6 mg/kg per d in the 0·1 % AA group; 261·8 and 874·0 mg/kg per d in the 0·5 % AA group; and 1075·1 and 3058·5 mg/kg per d in the 2·0 % AA group, respectively.
Morphological and morphometric analysis
Retinal histology showed no abnormal changes in the central and peripheral retina in vehicle-treated rats that consumed any of the four diets; the retinas contained more than ten layers of photoreceptor nuclei in the central retina (Fig. 2(a)) and more than eight layers of cells in the peripheral retina (data not shown). At 7 d after a single MNU injection, the outer nuclear layer and the photoreceptor layer in the central retina disappeared or were reduced to a few rows of photoreceptor cell nuclei in rats fed a basal diet or 0·1 % AA diet (Fig. 2(b)). The remaining photoreceptor nuclei were densely stained and contained clumped chromatin. In contrast, the 0·5 and 2·0 % AA diets rescued retinal damage. A waving array of retina was seen due to the focal disappearance of photoreceptor nuclei in the 0·5 % AA diet group, while the retina showed almost normal histology in the 2·0 % AA diet group. To further evaluate the effects of the AA diet on retinal thickness at the central retina, the photoreceptor cell ratio was calculated. MNU-treated rats that received the 0·5 and 2·0 % AA diets had a statistically significant increase in their photoreceptor ratio compared with basal diet-fed rats (basal diet, 19·4 %; 0·1 % AA diet, 32·0 %; 0·5 % AA diet, 46·7 %; 2·0 % AA diet, 43·6 %; Fig. 3(a)). MNU-induced retinal degeneration initiates from the central retina and progresses to the peripheral retina(Reference Yoshizawa and Tsubura17); thus, the peripheral retina in MNU-treated rats appeared normal in all the four diet groups. The morphometrical evaluation of the peripheral retina in MNU-treated rats showed that the photoreceptor ratio did not change in any of the diet groups. MNU-treated rats that received the 0·5 and 2·0 % AA diets had a statistically significant decrease in the retinal damage ratio compared with basal diet-fed rats (basal diet, 39·5 %; 0·1 % AA diet, 21·2 %; 0·5 % AA diet, 2·7 %; 2·0 % AA diet, 6·1 %; Fig. 3(b)).
Histology of the central retina in rats treated with (a) vehicle only or (b) 50 mg/kg of N-methyl-N-nitrosourea (MNU). Rats were fed the basal, 0·1, 0·5 and 2·0 % arachidonic acid (AA) diets. At 7 d after a single intraperitoneal injection of MNU, the outer nuclear layer and the photoreceptor layer disappeared in rats fed the basal diet and the 0·1 % AA diet (arrows). In contrast, the 0·5 and 2·0 % AA diets prevented the retinal damage (arrows). GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; PRL, photoreceptor cell layer; PEL, pigment epithelial cell layer. Haematoxylin and eosin staining, 200 × .
Photoreceptor cell ratio in (a) the central retina and peripheral retina 7 d after a single intraperitoneal injection of N-methyl-N-nitrosourea (MNU) into rats fed the basal, 0·1, 0·5 and 2·0 % arachidonic acid (AA) diets. Rats treated with 50 mg MNU/kg that received the 0·5 or 2·0 % AA diet had a statistically significant increase in their photoreceptor ratio at the central retina, compared with basal diet-fed rats. The index was calculated as: (outer retinal thickness/total retinal thickness) × 100. (b) Retinal damage ratio in MNU-treated rats fed the basal, 0·1, 0·5 and 2·0 % AA diets. Rats fed the 0·5 or 2·0 % AA diet had a statistically significant decrease in the retinal damage ratio compared with basal diet-fed rats. The index was evaluated as: (length of retina composed of less than four photoreceptor cells/whole retinal length) × 100. Values are means of four to eight rats in each treatment group, with their standard errors represented by vertical bars. ** Mean values were significantly different from those of the vehicle-treated group (P< 0·01).
Apoptotic index
At 24 h after the 50 mg MNU/kg treatment, TUNEL signals in the central retina of basal diet-fed rats appeared selectively in the photoreceptor cell nuclei; many pyknotic photoreceptor cell nuclei were TUNEL-positive without obvious nuclear destruction (Fig. 4(a)). TUNEL signals were detected much less frequently in the photoreceptor cell nuclei of the 0·5 and 2·0 % AA groups. The apoptotic indices in the central retina of MNU-treated rats that received the 0·5 and 2·0 % AA diets were significantly less (21·9 and 14·9, respectively) than the apoptotic index of MNU-treated rats that received the basal diet (37·6) (Fig. 4(b)). In the peripheral retina, the apoptotic indices of MNU-treated rats that received the 0·5 and 2·0 % AA diets were comparable with that of MNU-treated rats that received the basal diet.
Changes in the expression of terminal deoxynucleotidyl transferase-mediated dUTP digoxigenin nick-end labelling (TUNEL)-positive photoreceptor cells in the outer nuclear layer of the central retina. (a) At 24 h after a single intraperitoneal injection of N-methyl-N-nitrosourea (MNU) into rats fed the basal, 0·5 or 2·0 % arachidonic acid (AA) diet, the signals are seen in many photoreceptor cell nuclei of rats fed the basal diet. However, the signals are seen in only some photoreceptor cell nuclei of rats fed the 2·0 % AA diet. IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; PRL, photoreceptor cell layer. TUNEL staining, 400 × . The apoptotic index was evaluated in (b) the central retina and peripheral retina, respectively. Values are means of five rats, with their standard errors represented by vertical bars. Mean values were significantly different from those of the vehicle-treated group: * P< 0·05, ** P< 0·01.
Serum and retina fatty acid composition in N-methyl-N-nitrosourea-untreated rats
In the serum of vehicle-treated rats that were fed the basal, 0·1, 0·5 and 2·0 % AA diets, the AA compositions were 20·03, 22·66, 33·37 and 40·48 mol% of total fatty acids, and the AA levels were 0·37, 0·42, 0·62 and 0·74 μg/μl, respectively. The AA compositions and levels in 0·5 and 2·0 % AA diet-fed rats were significantly increased compared with those in rats fed a basal diet (Fig. 5). The serum AA:DHA ratio was significantly increased in 0·5 and 2·0 % AA diet-fed rats compared with that in rats fed the basal diet; the AA:DHA ratio was 3·92, 4·20., 8·86 and 18·00 in basal, 0·1, 0·5 and 2·0 % AA diet-fed rats, respectively. In contrast, in the retinas of vehicle-treated rats that were fed the basal, 0·1, 0·5 and 2·0 % AA diets, the AA compositions were 6·61, 7·58, 7·36 and 8·85 mol% of total fatty acid, and the AA levels were 1·60, 1·70, 1·78 and 2·18 μg/mg, respectively. The AA compositions and levels in 2·0 % AA diet-fed rats were significantly increased compared with those in rats fed the basal diet (Fig. 5). The AA:DHA ratio was significantly increased in rats fed the 2·0 % AA diet compared with that in rats fed a basal diet; the values were 0·46, 0·48, 0·54 and 0·63 in the basal, 0·1 %, 0·5 % and 2·0 % AA diet groups, respectively. In the basal, 0·1 %, 0·5 % and 2·0 % AA diet groups, the total fatty acid levels were 1·9, 1·7, 2·1 and 1·8 μg/μl in serum and 25·0, 23·1, 25·1 and 26·3 μg/mg in the retina, respectively (NS).
Fatty acid composition of the retinas and serum of vehicle-treated rats. In the serum of rats fed a 0·5 or 2·0 % arachidonic acid (AA) diet, (a) the AA composition of total fatty acids and (b) the AA:DHA ratio increased compared with those of basal diet-fed rats. In the retinas of rats fed a 2·0 % (AA) diet, (a) the AA composition and (b) the AA:DHA ratio increased compared with those of basal diet-fed rats. Values are means of four to eight rats in each treatment group, with their standard errors represented by vertical bars. Mean values were significantly different from those of the basal diet-fed group: * P< 0·05, ** P< 0·01.
Discussion
The present study examines the protective effects of dietary AA supplementation during the gestational, lactational and post-weaning periods against MNU-induced retinal degeneration in young rats. Morphological and morphometric evaluations indicate that a diet containing 0·5 or 2·0 % AA can counteract the effects of MNU and protect against MNU-induced retinotoxicity. Cell loss via apoptosis occurs in the course of human RP, as well as in mutant mouse models(Reference Hartong, Berson and Dryja9, Reference Rivas and Vecino14, Reference Sancho-Pelluz, Arango-Gonzalez and Kustermann15). Apoptosis has been considered as a final common pathway of photoreceptor cell death. In the present study, 0·5 and 2·0 % AA supplementation during prenatal development and early life inhibited the occurrence of apoptotic photoreceptor cell death 24 h after a single intraperitoneal injection of MNU.
AA and DHA are long-chain PUFA found in breast milk. Recently, it has been recommended that these PUFA be added to infant formula (0·2–0·5 % of total fatty acids)(Reference Hoffman, Boettcher and Diersen-Schade24). AA and DHA play a pivotal role in neurodevelopment and the response to neural injury in the neonatal stage(Reference Saste, Carver and Stockard5), and they are the major structural long-chain PUFA in retinal photoreceptor outer segments(Reference Neuringer and Dobbing25). The amounts of fatty acid in retinal tissue are modifiable by the composition of dietary fatty acid intake(Reference Hussein, Fedorova and Moriguchi20, Reference Suh, Wierzbicki and Lien26). The retinal levels of AA and/or DHA are lower in rats with rhodopsin mutation(Reference Anderson, Maude and McClellan27) and in rd mice(Reference Scott, Racz and Lolley28) with inherited retinal degeneration, and the plasma levels are lower in patients with human RP (Usher syndrome)(Reference Bazan, Scott and Reddy29). In the present study, the serum and retinal levels of AA and the ratio of AA:DHA increased in vehicle-treated rats fed the 0·5 or 2·0 % AA diet compared with rats fed a basal diet (containing < 0·01 % (w/w) AA). Retinal AA levels in the MNU-treated groups were not analysed, because different degrees of rod outer segment damage affected by an AA-rich diet may alter the levels and create misleading results. However, AA administered during early developmental stages may protect against MNU-induced photoreceptor cell degeneration in the rat retina.
The visual and cognitive development of humans is supported through 39 months of age by the consumption of infant formula supplemented with AA (0·43 % of total fatty acids) and DHA (0·12 % of total fatty acids) until the age of 1 year(Reference Auestad, Scott and Janowsky30). Visual maturation progression (a higher level of visual-evoked potential acuity) occurred in infants who received formulas supplemented with AA (0·72 % of total fatty acids) and DHA (0·36 % of total fatty acids) until 1 year of age(Reference Birch, Castaneda and Wheaton31). Consensus recommendations by the World Association of Perinatal Medicine and other societies specify that infant formula should contain DHA in the range of 0·2–0·5 % of total fatty acids and AA at a level that is at least equal to that of DHA(Reference Hoffman, Boettcher and Diersen-Schade24). Dietary DHA is effective in rescuing MNU-induced retinal degeneration in adult rats(Reference Moriguchi, Yuri and Yoshizawa32, Reference Tsubura, Yuri and Yoshizawa33). High DHA composition during the induction/signalling phase and/or the effector phase of retinal cell apoptosis can delay the onset of cell death and counteract the progression of retinal degeneration in rodents(Reference Moriguchi, Yoshizawa and Shikata34). In the present dietary dose regimen, the percentage of AA in total fatty acids and the AA:DHA ratio were extremely high (Table 1) compared with the recommended ratio in humans. In studies of retinal neurons in culture, DHA supplementation decreased photoreceptor cell death, whereas palmitic acid, oleic acid and AA were unable to decrease photoreceptor cell degeneration(Reference Moriguchi, Yoshizawa and Shikata34–Reference Rotstein, Aveldano and Barrantes36). However, in cultured brain neurons, AA shows biphasic actions at a narrow concentration range with the induction of cell death, on the one hand, and the promotion of cell survival and the enhancement of neurite extension, on the other hand(Reference Katsuki and Okuda37). The neurotoxic action is mediated by free radicals generated by AA metabolism, whereas the neurotrophic actions are exerted by AA itself(Reference Katsuki and Okuda37, Reference Kim, Akbar and Kim38). This evidence suggests that dietary AA supplementation may be beneficial as a potential means to delay the onset and/or progression of neural disease by the inhibition of neuronal cell death at narrow windows of susceptibility (in the developmental phase) for photoreceptor rescue. In the present study, there were no abnormal changes in the retinas or brains (data not shown) of vehicle-treated rats that were fed any of the AA-supplemented diets (0·1, 0·5 or 2·0 %) at the developmental stage. The 0·5 and 2·0 % AA diets or the 2·0 % AA diet significantly elevated serum and retinal AA levels as well as AA:DHA levels, respectively, and rescued photoreceptors from apoptosis.
In human subjects, AA supplementation by healthy adults appears to confer no toxicity or significant safety risk; daily dosages of 1500 mg for 50 d in the USA and 838 mg for 14 d in Japan have been well tolerated in clinical studies with no significant side effects(Reference Kusumoto, Ishikura and Kawashima39, Reference Nelson, Schmidt and Bartolini40). The recommended intake of AA in Japan is 24 mg/kg per d in adult human subjects(41). Thus, in the basal, 0·1, 0·5 and 2·0 % AA diets, the AA dose in the pregnancy period was 0·2, 3·2, 10·9 and 44·8 times higher than the recommended human dose, and the AA dose in the lactation period was 0·4, 10·1, 36·4 and 127·4 times higher than the recommended human dose, respectively. The average dietary AA intake was recently estimated to be 57 mg/d in healthy Canadian children between 4 and 7 years of age(Reference Lien and Clandinin42). The American College of Toxicology reviewed the safety and pharmacological assessments of AA, and found that exogenous AA appears to help prevent teratogenic effects in animal models and that a daily oral dose of 3000 mg AA/kg for 27 d shows no side effects in rats(43). Therefore, a higher dose of AA is tolerable and is needed for photoreceptor rescue as well as anti-teratogenic action in animals. Recently, female but not male newborn mice from C57BL/6 dams fed AA(+)/DHA(+) diets had a higher incidence of eye abnormalities (microphthalmia and corneal opacity)(Reference Maekawa, Iwayama and Watanabe44). However, a high incidence of eye abnormalities exists in both female and male C57BL/6 mice in historical control data(Reference Smith, Sundberg and Smith45). Therefore, further study is needed to determine whether AA exposure is the cause of sex-dependent eye abnormalities in C57BL/6 mice.
In conclusion, AA-rich diets (0·5 and 2·0 % AA) during the gestational, lactational and post-weaning periods prevent MNU-induced retinal degeneration in young rats via the inhibition of photoreceptor apoptosis. An AA-rich diet induces increased levels of retinal AA, probably followed by the promotion of retinal development and maturity. Several factors including AA itself in the retina may inhibit photoreceptor cell death via the apoptosis cascade in MNU-induced retinal degeneration. The induction/signalling phase and/or the effector phase of retinal cell apoptosis may delay the onset of cell death and counteract the progression of retinal degeneration in rodents fed an AA-rich diet, similarly to high DHA composition(Reference Moriguchi, Yuri and Yoshizawa32–Reference Moriguchi, Yoshizawa and Shikata34). Moreover, the detailed analysis of AA metabolites (cyclo-oxygenase and PG) in the retina of rats fed AA-rich diets may be necessary to understand the relationship to the rescue from MNU-induced retinal degeneration. The fatty acid composition of the prenatal and postnatal diet may affect the response to photoreceptor injury in human RP. Further studies of the photoreceptor apoptosis cascade are necessary to understand the detailed mechanisms by which AA prevents MNU-induced retinal degeneration.
Acknowledgements
This study was supported in part by a Grant-in Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (22591954) and Health and Labour Sciences Research Grants (H22-Shokuhin-Ippan-002). We thank Ms T. Akamatsu for her excellent technical assistance and Ms A. Shudo for manuscript preparation. The authors’ responsibilities were as follows: K. Y., T. S. and N. U. designed the study and provided technical expertise; K. Y. and M. K. conducted the study and histopathological analysis; A. H., N. O. and T. M. conducted and/or estimated the fatty acid analysis; K. Y., H. T. and A. T. analysed the data and wrote the manuscript. All authors read and approved the final manuscript. The authors declare that they have no competing interests.
