Elevated concentrations of organic and inorganic forms of iron in plant-based diets for channel catfish prevent anemia but damage liver and intestine, respectively, without impacting growth performance

We compared the effects of using inorganic and organic forms of iron in plant-based diets on catfish performance in a feeding trial with 6-g catfish fingerlings. The objective was to determine whether dietary iron in excess of known requirements negatively affected the fish. Five diets supplemented with 0 (basal), 125, 250 mg Fe/kg of either FeSO4 or iron methionine were formulated. Weight gain, feed conversion ratio, hepatosomatic index, and survival were similar among diets. Plasma and intestine iron concentration was similar among diets. Whole-body total lipid, protein, and dry matter were similar among diets, while ash content was higher in fish fed the basal diet. Total liver iron concentration was higher in fish fed diets supplemented with 250 mg Fe/kg in both iron forms than other diets. Hematological parameters were similar among treatments. Liver necrosis, inflammation, and vacuolization were highest in fish fed the diet supplemented with 250 mg Fe/kg from organic iron, followed by those fed diets with 250 mg Fe/kg from inorganic iron. Inorganic iron-supplemented diets caused more intestinal inflammation (increased inflammatory cells, villi swelling, thicker lamina propria) than the organic iron-supplemented diets or basal diet. Organic iron at 250 mg/kg resulted in a $0.143/kg increase in feed cost. Latent iron deficiency and initial signs of anemia developed in catfish fed the basal diet. Supplemental iron from either form prevented iron deficiency. Organic iron at 125 mg/kg optimized fish performance at a cost comparable to that of fish fed other diets, but without overt negative effects.


Introduction
Channel catfish (Ictalurus punctatus) production accounts for approximately 51% of total US finfish production and was valued at US$ 366.8 million annually in 2018 (USDA APHIS VS 1995; USDA NASS 2019). Commercial catfish production has declined in the past two decades due to competition from imports and increased feed costs (Hanson and Sites 2013;USDA NASS 2016). To restore industry competitiveness, alternative feed ingredients must be considered because feed cost is 50% or more of operating budgets (Engle and Stone 2014).
Soybean meal (dehulled, solvent extracted, 48% protein) is a widely preferred plant protein source in commercial catfish diets (Li et al. 2012). Soybean meal has a fairly well-balanced amino acid profile, is palatable to catfish, and contains lysine-the most limiting amino acid in catfish feeds (Dersjant-Li Abstract We compared the effects of using inorganic and organic forms of iron in plant-based diets on catfish performance in a feeding trial with 6-g catfish fingerlings. The objective was to determine whether dietary iron in excess of known requirements negatively affected the fish. Five diets supplemented with 0 (basal), 125, 250 mg Fe/kg of either FeSO 4 or iron methionine were formulated. Weight gain, feed conversion ratio, hepatosomatic index, and survival were similar among diets. Plasma and intestine iron concentration was similar among diets. Wholebody total lipid, protein, and dry matter were similar among diets, while ash content was higher in fish fed the basal diet. Total liver iron concentration was higher in fish fed diets supplemented with 250 mg Fe/ kg in both iron forms than other diets. Hematological parameters were similar among treatments. Liver necrosis, inflammation, and vacuolization were highest in fish fed the diet supplemented with 250 mg Fe/ kg from organic iron, followed by those fed diets with 250 mg Fe/kg from inorganic iron. Inorganic ironsupplemented diets caused more intestinal inflammation (increased inflammatory cells, villi swelling, thicker lamina propria) than the organic iron-supplemented diets or basal diet. Organic iron at 250 mg/kg 2002; NRC 2011). However, soybean meal contains anti-nutritional factors such as phytate that reduce the bioavailability of required minerals such as iron (Yuan et al. 2009;Hardy 2010;Kumar et al., 2012a).
Fish require iron for its role in cellular respiration (Lim et al. 1996;Davis and Gatlin 1996). Iron is a component of hemoglobin, which is responsible for oxygen transportation (Davis and Gatlin 1996;NRC 2011). Dietary iron deficiency or poor dietary iron digestibility can cause anemia in fish, because dietary iron is the major source of iron (Gatlin and Wilson 1986;NRC 1993;Lim et al. 1996;Evliyaoğlu et al. 2022). To prevent anemia, plant-based diets can be supplemented with organic forms of minerals, such as amino acid chelates (Suzuki et al. 1982; Davis and Gatlin 1996;Sun et al. 2015). These chelates (e.g., iron methionine) are absorbed actively in the gut along with other dietary amino acids and are less subject to binding by phytate compared to inorganic iron (e.g., FeSO 4 ) (Davis and Gatlin 1996;Chanda et al. 2015).
The results of Lim et al. (1996) indicated that organic and inorganic forms of iron were equally effective in supporting growth and hematological parameters in channel catfish. However, purified diets containing chemically defined ingredients were used, and the results could be quite different if these forms of iron were compared using typical commercial (practical) diets. Commercial diet composition has shifted significantly toward plant ingredients and away from animal ingredients (Li et al. 2010), reducing the amount of bioavailable forms of iron (Davis and Gatlin 1996). Current practical diets for catfish can include 95% or more plant ingredients. The main reason that organic minerals (including organic iron) have not been adopted by the catfish industry is their higher cost (current price US$ 3.9 per kg organic iron), along with a lack of data indicating their benefits when practical diets are used. Therefore, additional research on organic minerals in catfish is warranted.
This study was inspired by a current industry problem involving anemia outbreaks of unknown etiology in catfish (Lovell 1983: Camus et al. 2014. Despite the lack of supporting data, commercial producers have been using high-iron diets to treat the anemia. The dietary requirement of iron for optimal catfish growth and hematological parameters is 30 mg Fe/ kg diet (Gatlin and Wilson 1986). However, the highiron diets contain at least 10 times this amount. Overfortification with inorganic iron adds little cost to the diet, but excess iron is toxic-it acts as a pro-oxidant that can destroy other nutrients in the diet, and damage internal organs (Desjardins et al. 1987;Davis and Gatlin 1996;Sealey et al. 1997;Evliyaoğlu et al. 2022). Use of high-iron diets to mitigate anemia has become widespread in the catfish industry with little evidence to support the practice. Therefore, we assessed the effects of different concentrations and forms (inorganic and organic) of iron in plant-based diets on the growth performance, hematological parameters, tissue iron concentrations, and liver and intestinal histology of channel catfish.

Diet formulation and nutrient composition
Five plant-based catfish diets were formulated for a feeding trial to contain either inorganic iron from ferrous sulfate or organic iron from iron methionine at different levels of iron supplementation (0 mg Fe/ kg (basal), 125 mg Fe/kg, and 250 mg Fe/kg dry diet) ( Table 1).
All diets were formulated to meet the known nutrient requirements of catfish (NRC 2011) except for iron concentrations ( Table 2).
The basal diet contained 69.4 mg/kg diet intrinsic iron (Table 2). This concentration is approximately twice the published requirement for growth and absence of deficiency signs in catfish (Gatlin and Wilson 1986). Practical plant ingredients already contain high intrinsic iron levels (NRC 2011), so it was not possible to formulate an ironfree basal diet. The high supplementation levels of iron were tested to see if signs of iron toxicity would occur. All diets were formulated and made by a commercial feed manufacturing company (Ziegler Bros., Inc. Gardners, PA, USA). Diets were extruded as 2.5-mm floating pellets. Diets were analyzed for protein, dry matter, and ash content using standard methods (AOAC 2002), total lipid using Folch et al. (1957), and fiber using ANKOM Fiber Analysis (ANKOM200/220 fiber analyzer; ANKOM Technology, USA) following the manufacturer's protocol (Table 1). Dietary iron concentrations were analyzed using an Atomic Absorption Spectrometer (AAS) (Thermo Scientific iCE 3000 SERIES; USA) following the methodology described by Egnew et al. (2021) before the feeding trial. Feeding trial Juvenile channel catfish were obtained from the UAPB Aquaculture Research Station and transported to the fish nutrition wet lab, where they were acclimated to lab conditions for 2 weeks before the trial. During acclimation, fish were fed a commercial diet with 32% protein. The trial was conducted in eighteen 240-L tanks in a recirculating system supplied with dechlorinated municipal water. Each tank was supplied with a continuous water flow rate of approximately 2 L/min and constant aeration. The temperature was maintained between 27.5 and 30.0°C using thermostats. Salinity was maintained at 1-3 g/L by addition of common salt (NaCl) to counteract the toxic effects of nitrite in our recirculation system. Hardness of 80-120 mg/L/CaCO 3 was achieved by adding calcium carbonate (CaCl 2 ) whenever required. Before stocking, ten fish were randomly selected, euthanized with MS 222 (300 mg/L), and then stored at −70°C for whole-body proximate analysis. Fifteen fish with a mean individual initial weight of 6.1±0.2 g were stocked into each tank. Tanks were randomly assigned to one of the five diets with three replicates per diet. Fish were then fed twice daily to apparent satiation with their respective diets. The daily feed intake was quantified by subtracting the amount of feed remaining each day after feeding to satiation from the amount of the allocated feed. Mortalities occurring in the first 2 weeks of the trial were replaced with fish of similar size to compensate for stocking stress. Group weights of fish in each tank were measured every 2 weeks to monitor the growth and adjust the feed rations. The trial continued until a 500% increase in the fish growth for each diet was achieved (i.e., 10 weeks), and final fish weights and survival were determined. At this point, three fish were randomly collected from each tank, weighed, euthanized by MS 222 (300 mg/L), and then stored at −70°C for whole-body proximate analysis. The remaining fish were fed once daily to satiation for 2 weeks before samples for hematological assays and analysis of trace metals were collected. This waiting period was imposed to reduce the effects of sampling stress on hematological assays.

Water quality monitoring
Water temperature and mortalities were monitored and recorded daily. Measurement of pH, total ammonia nitrogen (TAN, salicylate/cyanurate method), hardness, dissolved oxygen, alkalinity, iron, and nitrite were conducted every week from three random tanks to ensure suitable conditions. The pH was measured using an Ultrabasic UB-10 pH meter (Denver Instrument, Bohemia, NY). Total ammonia and nitrite were measured using the Hach Test'N'Tube ® and NitiVer®3 test (Hach Company, Loveland, CO), and un-ionized ammonia values were then calculated (Trussell 1972). Hardness values were obtained using the USEPA Manver 2 Buret Titration method (Hach method 8226), iron using the FerroZine Method (HACH method 8147) adapted from Stookey (1970), and salinity using an Aquafauna salinity refractometer (Aquatic Eco-systems, Apopka, FL).

Growth performance
Fish weight gain, feed conversion ratio (FCR), and survival were computed for all treatments using the following formulae: Mean weight gain (g) = Mean final weight -mean initial weight SGR = (lnWf−lnWi×100) t , Where: lnWf = the natural logarithm of the final weight; lnWi = the natural logarithm of the initial weight; t = time (days) between lnWf and lnWi Proximate analysis of the whole body was conducted on three fish per tank that were collected at the end of the trials (described in the growth trial section). For whole-body analysis, fish samples were ground using an electric meat grinder (Harbor Freight Tools, UL, 99598), freeze dried, and analyzed for total lipids, protein, dry matter, and ash as described for the diets. An additional three fish per tank were randomly selected, weighed, and anesthetized in MS222 (100 mg/L) to collect blood for hematology and iron analysis (described in hematology and total plasma sections). Then, the bled fish were euthanized in tricaine methane sulfonate (MS-222) at 300 mg/L Weight gain (%) = total weight gain (g) initial weight (g) × 100 FCR = Mean feed intake (g) Mean weight gain (g) Survival (%) = initial fish number − number of dead fish initial fish number × 100 and dissected to extract and weigh the livers. The hepatosomatic index (HSI) was then calculated from the following: The liver samples were preserved for iron analysis. The same fish were used to collect muscle and intestine tissues preserved for iron assays (described in total plasma iron and tissue iron analysis sections).

Hematological analysis
Blood samples were collected from the same three fish per tank from which tissues were extracted for iron assays (as described in the tissue iron analysis section). The fish were first anesthetized in MS222 (100 mg/L). Blood was drawn from the anesthetized fish by caudal puncture using a heparinized syringe placed in Eppendorf tubes. The heparinized blood was used to determine hematocrit (Hk) by centrifuging the heparinized blood samples collected in the heparinized microhematocrit capillary tubes at 3500 x g for 10 min and hemoglobin (Hb) following the methods of Houston (1990). The mean corpuscular hemoglobin content (MCHC) was also calculated from MCHC = Hb g dL Hk . Plasma samples were collected and used for blood iron analysis (as described in the total plasma iron section).

Total plasma iron
Iron content in the plasma was determined using an Atomic Absorption Spectrometer (AAS) (Thermo Scientific iCE 3000 series; USA) following protocols described by Egnew et al. (2021). Plasma samples were obtained from heparinized blood collected as described in the previous section by centrifuging at 800 × g for 15 min and stored at −40°C. Briefly, 400 μL of high purity nitric acid (HNO 3 ) was added to 200 μL of plasma sample, and 40 μL of hydrogen peroxide (H 2 O 2 ) was added to the mixture. The mixture was digested at 105°C for 30 min in a hot block (Environmental Express HotBlock® SC154-240 Digestion System, 54-Well, 50 mL; 240 VAC). The digested samples were diluted with distilled water and analyzed for iron concentration using the AAS with HSI = liver weight body weight × 100.
deuterium lamp background correction and calibration using iron standard solutions (CPI International, CA, USA).

Tissue iron analysis
Total iron concentration in the intestine, liver and muscle tissues was determined using the AAS (Thermo Scientific iCE 3000 SERIES; USA) following a slight modification in the sample digestion protocol . The tissue samples were collected as described in the growth performance section and stored at −40°C until the iron analysis was conducted. The tissue samples were dried, weighed, and digested (acid digestion) in a hot block (Environmental Express HotBlock® SC154-240 Digestion System, 54-Well, 50 mL; 240 VAC) before analysis. Briefly, 4 mL of HNO 3 were added to 0.4 g of dried sample. Samples in HNO 3 were allowed to digest at room temperature under a fume-hood for 12 h and then digested at 115°C in a hot block for 30 min. Then, 100-200 μL of H 2 O 2 was added to the samples, and further digested at 115°C for 30 min. Digested samples were allowed to cool and diluted with deionized water. Iron concentrations (μg/L) in the digested samples were then read in the AAS as per the total plasma iron estimation (described above).

Tissue collection and fixation
Two additional fish per tank were euthanized by MS 222 (300 mg/L) and dissected to extract tissues (small intestine and liver) for histological analysis.

Tissue processing and histological analysis
Histological analysis was conducted using procedures described by Romano et al. (2021). Briefly, extracted tissues were immediately fixed in Bouin's solution for 18 h and then preserved in 70% ethanol. Then, the tissues were processed in a tissue processor through a series of varying ethanol concentrations starting with 70% (v/v) to allow dehydration, followed by clearing in xylene. The processed tissue samples were embedded in paraffin wax in metal molds and allowed to harden into blocks which were then sectioned using a rotary microtome (HM 294 Fish Physiol Biochem (2023) 49:289-305 340E, Thermo Scientific) into 5-μm sections. The slices were then mounted on microscope slides, stained with hematoxylin and eosin and analyzed in a light microscope (Leica DM 2500 LED) for histopathological changes. Leica Application Suite software was used to measure villi width and length before calibration, after taking the pictures. A semiquantitative histological scoring system was used to estimate the severity of the histopathological change present, i.e., 1 (normal), up to 4 (severe) for liver, and 1 (normal) up to 3 (severe) for intestine (O'Connell 1976;Hartviksen et al. 2014;Palma et al. 2021).

The estimated economic impact of dietary variables
Cost-of-gain analysis was conducted to determine the differences in feed cost per unit weight gained by the fish due to dietary iron supplementation of either organic or inorganic iron. Cost of gain was calculated from the formula (feed cost of gain = feed cost ($)/weight gain (kg)). A partial budget analysis was not conducted since there was no significant change in performance and production yield metrics (i.e., weight gain, survival, and FCR) brought about by either organic iron or inorganic iron.

Statistics
All percentage data (hematocrit, proximate composition, percentage survival) were arcsine transformed prior to analysis. Analysis of variance (ANOVA) was used to analyze data with iron form and iron concentration as the fixed effects. The interaction between the two iron forms (inorganic and organic iron) and their effect at the different concentrations was considered significant at p ≤0.05. Differences among treatment means were determined using Tukey's post hoc test and considered significant at p≤0.05 in R software version 3.5.3.

Growth, feed efficiency, and survival
Mean individual weight gain, specific growth rate, feed intake, FCR, and survival did not differ among diets (Table 3). There was no significant interaction between iron forms and iron concentration on any of these parameters. Survival in all dietary treatments was similar and uniformly high (95.6-100%) ( Table 3). Whole-body proximate analysis and hepatosomatic index (HSI) No differences were observed in hepatosomatic index, total lipid, or dry matter of whole-body among diets (Table 4). Protein content was similar among fish fed the basal diets and diets with 250 mg Fe/kg (both forms). However, whole-body protein of fish fed the diet with 125 mg FeSO 4 /kg was higher than that of fish fed the diet with 125 mg FeM/kg. Ash content was higher in fish fed the basal diet not supplemented with iron compared to those fed the iron-supplemented diets.
Total iron concentration of plasma, intestine, liver, and muscle Total iron concentration in the plasma and intestine were similar among diets (Table 5).
Total liver iron concentration was higher in fish fed diets supplemented with 250 mg Fe/kg in both iron forms than in the fish fed with either the basal diet or the 125 mg Fe/kg diets (Table 5). Iron concentration in muscle was below detectable limits (0.025 mg/L).
Hematological parameters-hematocrit, hemoglobin, and mean corpuscular hemoglobin concentration (MCHC) There were no differences in hemoglobin concentration, hematocrit, or mean corpuscular hemoglobin concentration (MCHC) of fish among diets ( Table 6).

Histopathology of the liver and intestine
Liver Iron-supplemented diets with 250 mg Fe/kg caused severe hepatic necrosis, inflammation (presence of leukocytes), vacuolization, and vessel congestion than in all other groups ( Figs. 1 and 2).
Iron-supplemented diets with 125 mg Fe/kg and the basal diet did not cause hepatic necrosis, inflammation, vacuolization, or congested blood vessels in fish. In contrast, the organic iron-supplemented diet at 250 mg Fe/kg caused liver necrosis, inflammation, and vessel congestion in the fish compared to the inorganic iron diet at 250 mg Fe/kg ( Figs. 1 and 2).   Inorganic iron-supplemented diets caused more severe intestinal inflammation in the fish compared to the organic iron-supplemented diets.
Iron-supplemented diets from either source caused an increase in the intestinal goblet cells in fish compared to the basal diet (Figs. 3 and 4). There was no change in catfish growth performance (i.e., weight gain, survival, and FCR) brought about by either organic iron or inorganic iron ( Table 7). Cost of gain increased by 0.143 $/kg for fish-fed diets with supplemental organic iron at 250 mg/kg diet, compared to the basal diet (Table 7). Dietary iron supplementation with organic iron at 125 mg/kg diet, and inorganic iron at a maximum of 250 mg/kg diet, added minimal cost compared to the basal diet. Dietary organic iron supplementation of 250 mg/kg diet produced the highest increase in fish feeding cost per unit of weight gain.

Discussion
Growth (weight gain), FCR, and survival did not differ among diets in channel catfish-fed plant-based diets supplemented with either inorganic iron or organic iron up to 250 mg/kg diet. Fish weight gain reportedly decreased in channel catfish fingerlings fed a purified diet with no supplemental iron compared to those fed iron-supplemented purified diets (20-300 mg Fe/kg diet) (Gatlin and Wilson 1986;Lim et al. 1996;Lim and Klesius 1997;Sealey et al. 1997). However, in this study with plant-based practical diets, fish weight gain during the trial period was not affected by dietary iron supplementation. Gatlin and Wilson (1986) also reported a significant improvement in feed efficiency of fish-fed purified diets with supplemental dietary iron (up to 50 mg/kg) compared to those fed an unsupplemented diet. However, in this study, feed conversion was not affected by dietary iron supplementation. Lim and Klesius (1997) reported low survival of juvenile channel catfish fed an iron-deficient-purified diet. However, survival of catfish in this study was similarly high among diets. Similar results were obtained by Gatlin and Wilson (1986) and Lim et al. (1996). Differences in the results of earlier studies versus the present study are most likely due to the high levels of intrinsic iron in the practical plant ingredients (69.4 mg/kg diet) used in this study. The intrinsic iron concentration in the basal diet exceeds the iron requirement reported for catfish growth and absence of deficiency signs (30 mg/kg diet; Gatlin and Wilson 1986). It was not possible to formulate a practical diet with no intrinsic iron using conventional ingredients. However, iron in plant materials is usually present as non-heme iron and is always bound to indigestible complexes (anti-nutritional factors) such as phytate. Minerals bound by phytate have low bioavailability to fish (Francis et al. 2001;Kumar et al. 2012a,b). Organic iron from iron methionine reportedly promotes better growth and feed efficiency in fish (Suzuki et al. 1982;Lim et al. 2000). However, no interaction between iron concentration and iron form was observed in this study or that of Lim et al. (1996), indicating that the supplemental iron source (i.e., ferrous sulfate or iron methionine) had little impact on catfish growth performance. Weight gain and feed efficiency of juvenile cobia (Rachycentron canadum) fed semi-purified diets with either ferrous iron or iron methionine were also similar (Qiao et al. 2013). Additional studies with other species such as yellowtail (Seriola quinqueradiata) (Ikeda et al. 1973), common carp (Cyprinus carpio) (Sakamoto and Yone 1978a), red sea bream (Pagrus major) (Sakamoto and Yone 1978b), Atlantic salmon (Salmo salar) (Andersen et al. 1996), juvenile gibel carp (Carassius gibelio) (Lei et al. 2009), and adult GIFT tilapia (Oreochromis niloticus) (Wen et al. 2019), also indicated that dietary iron supplementation did not affect growth and feed utilization.
In contrast, growth and feed utilization improved significantly in eel fed a diet supplemented with organic iron (Suzuki et al. 1982), Nile tilapia (Oreochromis niloticus)-fed purified diets supplemented with inorganic iron (El-Serafy et al. 2007), and rainbow trout (Oncorhyncus mykiss) -fed plant-based diets supplemented with inorganic iron (Evliyaoğlu et al. 2022). Differences in results of feeding trials with fish are often due to differences in experimental conditions, including variation in ingredient composition of diets.
Hepatosomatic index (HSI), total lipid, ash, and dry matter content of catfish were similar regardless of the tested dietary iron concentration. Similar findings were reported in rainbow trout (O. mykiss) fed semi-purified diets supplemented with inorganic iron (Carriquiriborde et al. 2004), and Atlantic salmon (Salmo salar) (Sutton et al. 2006) fed practical diets supplemented with inorganic iron. However, in the current study, the ash content was higher in catfish fed the basal diet compared to diets supplemented with iron. This result seems counter-intuitive, but at very high levels dietary iron can compete with other divalent minerals such as zinc, calcium, and cadmium for similar binding sites during intestinal absorption (Whitehead et al. 1996;Andersen et al. 1997;Lonnerdal 2010;Kwong et al. 2011;Lall and Kaushik 2021). Carriquiriborde et al. (2004) reported that manganese, copper, and zinc in rainbow trout (Oncorhynchus mykiss) tissues decreased with an increase in dietary supplemental iron in semi-purified diets (33, 175, and 1975 mg Fe/kg). In this study with catfish, relatively high concentrations of iron were present in the supplemented diets (125 mg/kg minimum, which is about four times higher than the published requirement (30 mg/kg) (Gatlin and Wilson 1986). Since ash content represents total body mineral content, interruption of mineral absorption might result in lower ash content.
In this study, whole-body protein was not affected by dietary supplemental iron. The low protein observed in fish fed the organic iron diet with 125 mg/kg could not be attributed to iron since there was no strong trend. Carriquiriborde et al. (2004) reported that dietary supplemental iron did not affect wholebody protein in rainbow trout (Oncorhynchus mykiss). Sutton et al. (2006) also observed that whole-body protein in Atlantic salmon (Salmo salar) was not affected by dietary supplemental iron in fish fed practical diets.
In the current study, catfish liver iron concentration increased with increasing supplemental dietary iron concentration. However, liver iron concentration in fishfed diets supplemented with 250 mg/kg iron from both sources was higher than those fed other diets. Liver iron concentration of fish fed the basal diet was 23 to 50% lower than that of those fed diets supplemented with 125 mg/kg diet, though concentrations did not differ statistically. Similar findings were reported by Lim et al. (2000) in channel catfishfed egg-white-based diets supplemented with organic iron (iron methionine). However, excessive supplemental dietary iron from inorganic iron (ferrous iron) did not cause an increase in hepatic iron in African catfish (Clarias gariepinus) (Baker et al. 1997) or in Atlantic salmon (Salmo salar) (Andersen et al. 1996;Andersen et al. 1998). Lim et al. (2000) and Camus et al. (2014) showed that liver iron concentration of less than 40 μg/g in channel catfish could indicate iron deficiency and microcytic anemia. In our study, liver iron concentration in catfish fed the basal diet was 33.9 μg/g. This concentration was consistent with signs of developing microcytic anemia, as documented in other fish species. For instance, when Atlantic salmon had a liver iron concentration of 30 μg/g, they could not sustain red blood cell synthesis and maintain or increase hemoglobin concentrations (Andersen et al. 1998). Microcytic anemia can be caused by dietary iron deficiency (i.e., low iron content in diets) or poor dietary iron digestibility (Massey 1992; Lim et al. 2000). In fact, several studies have shown that anemia could be induced in channel catfish-fed iron-deficient-purified diets (Gatlin and Wilson 1986;Lim et al. 1996;Lim et al. 2000). Irondeficiency-induced anemia also occurred in other fish species-fed iron-deficient-purified diets, including brook trout (Kawatsu 1972), yellowtail (Ikeda et al. 1973), and red sea bream (Sakamoto and Yone 1978b).
Hemoglobin concentration, hematocrit, and mean corpuscular hemoglobin concentration in catfish were not affected by dietary iron supplementation in this study. Hematological values decreased in juvenile channel catfish-fed iron-deficient-purified diets in other studies (Gatlin and Wilson 1986;Lim et al. 1996;Lim et al. 2000). Nevertheless, Camus et al. (2014) reported a serum iron content of 35.2 μg/dL in catfish with anemia. Serum iron concentrations in catfish fed the basal diet in this study were much lower (7.7 μg/dL). The lack of statistical differences in serum iron levels might have been due to high variability in the data. Gatlin and Wilson (1986) and Lim et al. (1996) reported a decrease in hematological values (development of microcytic anemia) in channel catfish fingerlings fed the iron-deficient-purified diet. Andersen et al. (1996) also reported a significant decrease in hematological values in Atlantic salmon (Salmo salar) fed the purified diet that was not supplemented with iron.
Channel catfish-fed diets supplemented with 250 mg Fe/kg developed severe hepatic necrosis, inflammation (presence of leukocytes-predominantly leukocytes), vacuolization, broken pancreatic islets, broken vessels, and vessel congestion compared with few or none of these pathologies in those fed diets with less iron. Hepatic iron concentrations in this study were also highest in fish-fed diets with 250 mg Fe/kg (134.7 μg/g dry weight). These findings indicate iron toxicity that was characterized by ferroptosis of the hepatocytes from liver iron overload in fish fed the diets supplemented with 250 mg Fe/kg (Wang et al. 2017). However, at 250 mg Fe/kg, the organic ironsupplemented diet caused more severe pathological effects in the fish liver than those fed the inorganic iron-supplemented diet. Organic iron is absorbed by active transport along with the amino acid bound to it in the chelate. Therefore, iron can be absorbed into the enterocytes in larger quantities than inorganic iron (FeSO 4 ) at the same dietary level (Chanda et al. 2015;Sun et al. 2015). When more iron is absorbed into the enterocytes and is readily available for metabolic needs, it is less likely that the fish will develop anemia from iron deficiency (Chanda et al. 2015).
When dietary iron is excessive, the increased quantities of absorbed iron may result in iron overload in the blood and organs-especially the liver. Such hepatic iron overload induces oxidative stress, which can result in liver damage (Desjardins et al. 1987;Liu et al. 2013;Smith et al. 2014). Both Desjardins et al. (1987 and Carriquiriborde et al. (2004) reported that excessive dietary iron induced unique histopathological changes in the liver of rainbow trout (Oncorhynchus mykiss). In channel catfish (Ictalurus punctatus) (Yadav et al. 2020) and Atlantic salmon (Salmo salar) (Valenzuela-Muñoz et al. 2020), tissue iron overload-induced pathological abnormalities in the liver.
Supplemental dietary iron also induced intestinal histological modifications in rainbow trout 302 Fish Physiol Biochem (2023) 49:289-305 (Oncorhynchus mykiss)-fed semi-purified diets (Desjardins et al. 1987;Carriquiriborde et al. 2004) and practical diets (Evliyaoğlu et al. 2022). Intestinal histological changes in catfish were characterized by increased villi and lamina propria width as well as a higher prevalence of goblet cells in fish-fed diets supplemented with inorganic iron compared to those fed the basal diet. However, the intestines had substantially less severe changes in fish-fed diets supplemented with organic iron and the basal diet, except for the goblet cells. The latter increased in fish-fed diets with supplemental organic iron compared to those fed the basal diet indicating a potential protective response because goblet cells are responsible for mucus secretion. These differences indicate that supplemental dietary iron negatively affected the gastrointestinal health of the fish. However, diets supplemented with inorganic iron caused more severe intestinal inflammation in channel catfish (characterized by the presence of leukocytes) compared to fish-fed diets with organic iron or no supplemental iron. Iron in the inorganic form (including salts such as ferrous sulfate) is free and in a highly reactive oxidative state (ferrous iron) (Davis and Gatlin 1996). Inorganic iron reacts with dietary lipids, causing iron-induced peroxidation (Desjardins et al. 1987;Sutton et al. 2006). Peroxidation induced intestinal histopathological changes, inflammation, intestinal oxidative stress, and reduced intestinal function in Atlantic salmon (Salmo salar) (Sutton et al. 2006), and rainbow trout (Oncorhynchus mykiss)-fed semi-purified diets (Desjardins et al. 1987;Carriquiriborde et al. 2004) and rainbow trout-fed plant-based diets (Evliyaoğlu et al. 2022).
The cost of gain increased by 0.143 $/kg for channel catfish-fed diets supplemented with organic iron at 250 mg/kg diet, compared to those fed other diets. This result indicates that dietary organic iron supplemented at 250 mg/kg resulted in a unit increase in fish feeding cost per unit weight gain. The cost of gain was similar for fish-fed diets supplemented with organic iron at 125 mg/kg diet, inorganic iron, and the unsupplemented basal diet. Considering other results, inorganic iron at both concentrations (i.e., 125 and 250 mg/kg diets) induced signs of toxicity in the fish intestine, and liver (i.e., 250 mg/kg diet). Organic iron at 250 mg/kg induced signs of toxicity in fish liver. The basal diet-induced signs of iron depletion and emerging microcytic anemia in the fish. However, dietary supplementation of organic iron at 125 mg/kg caused no apparent deleterious effects and improved the overall iron status of the fish.

Conclusion
The diet supplemented with 125 mg organic iron/kg optimized fish performance at a cost comparable to that of fish-fed other diets, but without any apparent negative effects.
It should be noted that these results were obtained under fixed, laboratory conditions designed to support rapid fish growth. In commercial catfish production in earthen ponds, there is inherent variability in temperature, dissolved oxygen, and other environmental variables that affect feed intake and growth. Anemia outbreaks sometimes occur during seasonal transitions, when temperature, feed intake, and growth are inconsistent. Under such conditions, using high-iron diets may not produce negative effects. In fact, using higher iron diets and/or use of the organic form could even be beneficial when water temperature and feed intake are low (typically, below 22°C). Therefore, further studies are needed to establish the minimum and maximum dietary organic iron concentrations to optimize performance of channel catfish-fed plantbased diets under different environmental conditions to help minimize outbreaks of anemia during catfish production.