The indoor management and maintenance of HSCs was evaluated in a pair of RAS at a salinity of 19.5–22.0‰. Each RAS was equipped with two holding tanks (4 ft × 6 ft × 1 ft), a biofiltration tank with K1 media (Pentair, Minneapolis, MN, United States), and a clarifying (solids separation) tank (Figure 2). Ambient air supplied oxygen to the water via a piston air pump with supplemental air stones positioned in each tank (Nelson and Pade, Montello, WI, United States). Water was recirculated through biological filtration tanks, allowing for water purification, hygiene, and disease prevention. Water parameters were measured and adjusted as needed six times per week over the course of the study.

HSCs were stocked at a density of one animal per 3 ft². An isolated circular resin tank (diameter: 3.25 ft) equipped with a solids filter was used as a “hospital” tank to quarantine HSCs prior to RAS introduction and allow for recovery after catheter surgery (see Catheter Implantation section). HSCs (n = 24) were procured (Dynasty Marine Associates, Marathon, FL, United States) or acquired and maintained in accordance with collection and aquaculture operation permits (North Carolina Department of Environmental Quality, Division of Marine Fisheries; #1946771 and #1947050). A health assessment was recorded for each HSC upon intake (i.e., sex, weight, and carapace length/width, as well as appendage, carapace, eye, and mouth evaluation). HSCs sourced directly from the wild (“wild type”) were used as controls for some studies, from which hemolymph extracts were collected within 1 h of habitat removal to establish baselines.

In order to ensure HSC well-being and optimal hemolymph parameters, diet and feeding rates were established over a trial period of 6 weeks prior to the initiation of the study. Feed rates, dietary protein (DP), nutrient composition, and daily energy requirements were determined for natural feed sources and a commercially manufactured aquafeed (Classic Brood 46/12 Complete Feed, Skretting, Tooele, UT, United States). Initial observations revealed that HSCs fed at different rates per day (expressed as a percentage of the b/w of the HSC) depending on the diet (Tinker-Kulberg et al., under review). In this study, the nutrient composition of the selected commercial aquafeed after modification with gelatin (5% w/v) included protein (23%), lipids (6%), carbohydrates (2%), and moisture content (50%). Gelatin served as a binding agent to provide rigidity, extended water stability and nutrient retention; thus eliminating buoyancy (for sinking density) and affording HSCs time to forage on the tank floor to accommodate characteristic bottom scavenging. A daily feed rate of 3% of their b/w sustained (or increased) HSC weight (∼1.2% change on average) and serum protein levels (above 5.0 mg/ml) over the course of the study, with an associated survival rate of 100%. Any feed not consumed after 8 h was removed. Sodium bicarbonate was added to the RAS at 0.12 g per 1.0 g of feed to maintain alkalinity and pH.

The catheter was composed of a radiopaque tube and a valve. For implantation, each HSC was gently immobilized, and the book gills were bathed with saltwater. Protective draping surrounded access to the pericardial membrane, and the region was prepped with 70% ethanol and betadine. An 18-gauge needle on the intravascular system was inserted at a 45° angle toward the prosoma (cephalothorax). Once the appearance of hemolymph was observed in the catheter, the guidewire was inserted and the device positioned. The needle was then removed; the site was cleaned; and a sterile, quick-drying topical tissue adhesive (Vetbond^TM, 3M, St. Paul, MN, United States) was applied. The catheter was capped with a sterile luer-lock, and it was affixed to the shell of the HSC in the dorsal prosoma region. Postoperative HSCs were monitored during recovery in the hospital tank. Using the implanted catheter, hemolymph was extracted for an initial health evaluation after 7 days, and the HSCs were reintegrated into the RAS cohort.

Hemolymph was collected via previously established methods for both the implanted and control cohorts (Coates et al., 2012; Anderson et al., 2013). Materials were rendered pyrogen-free by cleaning and overnight soaking in 1% E-Toxa-Clean(R) Solution (Sigma-Aldrich, St. Louis, MO, United States), heating at 200°C for 3 h, and rinsing with endotoxin-free water. For the control group without catheter insertion, the pericardial membrane was cleaned twice with 70% ethanol prior to hemolymph collection with a 22-gauge needle (Becton Dickson, NJ, United States) using a standard SOP (Armstrong and Conrad, 2008); for the implanted HSCs, it was collected using the catheters. An aliquot of the hemolymph was diluted (1:1) in a prechilled mixture of N-ethylmaleimide (0.125%), NaCl (3%), and Tris-HCl (0.5 M, pH 7.5) for amebocyte density analysis (Armstrong and Conrad, 2008; Coates et al., 2012; see Hemocyanin and Amebocyte Density section). Sterile tubes containing the balance of each harvest were centrifuged (1,000-rcf/5-min), and the supernatant was removed. Any samples that appeared clotted after centrifugation were not processed or analyzed. Amebocyte pellets were then washed with endotoxin-free NaCl (3% v/v), and both fractions were stored at −70°C pending analysis and LAL extraction.

Over the course of this research, to monitor the health of individual HSCs, 1 ml of hemolymph was extracted and analyzed for Hc concentration, amebocyte density, and LAL activity. For hemolymph extractions (e.g., 10%), blood volume was calculated as 25% (b/w), and the density was estimated at 0.00104 kg/ml (Hurton et al., 2005).

Hc and amebocyte density were measured throughout the study as an indicator of HSC health. Total serum protein concentration was used as a proxy for Hc by diluting the hemolymph supernatant in Tris-HCl (0.1 M, pH 7.5) and measuring absorbance at 280 nm on a UV-Vis Spectrometer (Cary 6000i, Agilent Technologies). Dioxygen-bound Hc (Oxy-Hc) was likewise measured at 340 nm. The total protein concentration was calculated according to Nickerson and Van Holde (1971), using the absorption coefficient A₂₈₀ = 1.39 mg^–1 ml^–1 cm^–1 for Hc, and A₃₅₀ = 0.223 mg^–1 ml^–1 cm^–1 for Oxy-Hc. The percentage of Hc that was oxygen bound was determined using the ratio of Oxy-Hc (mg/ml) to Hc (mg/ml) × 100.

A 10 μl volume of the diluted aliquot was placed in a hemocytometer and analyzed microscopically. Since amebocytes are susceptible to rapid degranulation, a digital camera was used to record bright-field images from the hemocytometer, and counts were processed off-line using the ImageJ software (National Institutes of Health, Bethesda, MD, United States).

LAL activity was assessed throughout the study as an indicator of HSC health and to investigate the effects of aquaculture on quality. LAL was prepared by thawing frozen amebocyte pellets on ice and then lysing them in pyrogen-free water at a 1:1 ratio of original hemolymph collection volume under gentle agitation over night at 4°C. Next, a chloroform extraction (1:1 v/v) of the lysate was performed for total protein measurement using a BCA kit (Pierce, Thermo Fisher Scientific, Waltham, MA, United States). All LAL extracts were aliquoted and frozen at −80°C pending analysis (no loss in LAL activity was observed in frozen samples for up to 6 months). Equal protein amounts of the LAL from aquaculture HSCs and commercial kits (E-Toxate^TM, Sigma-Aldrich, St. Louis, MO, United States) were diluted to a total volume of 100 μl and tested for their respective and relative activities. All evaluated LAL samples were subjected to a single freeze–thaw cycle to preserve activity.

LAL activity was evaluated turbidimetrically. The commercial kit served as a control, and assays were performed following the manufacturer’s instructions (Sigma-Aldrich, St. Louis, MO, United States). Aquaculture and commercial LAL (100 μl) samples were gently mixed with 100 μl of standard endotoxin solution (final concentration range: 0–50 EU/ml) in a pyrogen-free, 96-well microplate and incubated at 37°C for 1 h. The conversion of coagulogen to coagulin was measured at 340 nm on a microplate reader (BioTek 800^TM, BioTek Instruments, Winooski, VT, United States). The turbidity of the blank (endotoxin-free lysate) was subtracted from test values, and the relationship between endotoxin concentration and clot formation was measured against a standard curve derived from commercial LAL.

Equal amounts of amebocyte total protein isolated from the individual HSCs were pooled and analyzed for LAL activity in triplicate. A turbidimetric LAL assay was performed by incubating a standard endotoxin solution (50 EU/ml, Sigma Aldrich, St. Louis, MO, United States) with different concentrations of LAL (0–300 μg). Hemolymph extracted from wild-caught HSCs (“wild type”) was collected within 1 h of capture, from which LAL was processed and frozen at −80°C pending analysis and then used as a control; results were also compared to a commercial LAL kit (E-Toxate^TM, Sigma Aldrich, St. Louis, MO, United States). Both controls were diluted to the same concentration as the aquaculture LAL.

Purified protein extracts (10 μg) from commercial, aquaculture, and wild-type LAL were subjected to SDS-PAGE in the presence of β-mercaptoethanol. The total protein concentration of each sample was determined with a BCA assay. Each sample was diluted in a 4 × NuPAGE^TM loading buffer (Invitrogen, Carlsbad, CA, United States) containing 2.5% β-mercaptoethanol and heated to 85°C for 3 min. Electrophoresis was performed on 8% tris-glycine gels (Novec^TM Wedge Well, Invitrogen, Carlsbad, CA, United States) and stained with Coomassie Brilliant Blue R-250 (Bio-Rad, Hercules, CA, United States). A SeeBlue^® Plus2 Pre-stained Protein Standard (Invitrogen, Carlsbad, CA, United States) was used as a molecular weight marker.

To test the ability for LAL to detect endotoxins (LPS) in human blood, presumably endotoxin-negative human blood was freshly drawn from healthy volunteers using citrated collection tubes (3.2% buffered sodium citrate to inhibit clot formation). Aliquots were then spiked with various concentrations of LPS (Escherichia coli 055:B5, World Health Organization [WHO], 2012) or left neat as negative controls and incubated at room temperature for 15 min (to allow unbound LPS to disperse so as to mimic circulating endotoxins in patient specimens). Before LAL analysis, the samples were centrifuged; aliquots were removed and chemically treated with a matching volume of endotoxin-free reagents (to that of the initial volume) to remove blood components known to interact with LPS and inhibit LAL activation. LAL assays were then performed as described above and demonstrated consistent spiked bacterial concentrations.

To evaluate detection of intact bacteria in human blood, exponential growth of E. coli (Turbo Competent) was selectively achieved in LB broth containing 100 μg/ml of ampicillin to an optical density of 0.35. The culture was centrifuged to pellet the bacteria and washed twice to remove unbound LPS. A sample containing only the LB growth medium in the absence of bacteria was also processed in parallel, to account for any LPS contamination (negative “broth” control). Samples were then resuspended in isotonic buffer (0.9% NaCl), and a bacterial count (CFU/ml) was performed by plating serial dilutions on LB agar plates containing 100 μg/ml of ampicillin. Blood samples were then spiked with equal volumes of bacteria at different concentrations or with the negative “broth” control and incubated at RT for 15 min. All samples were chemically treated as with the LPS-spiked blood samples and then frozen at −70°C for 1 h to destabilize the E. coli cellular membranes. Samples were then thawed at RT, and LAL assays were performed.

LPS-spiked blood (0–50 EU/ml), E. coli-spiked blood (0–5.0 × 10⁵ CFU/ml), and controls were incubated with LAL at 37°C for 1 h at a 1:1 dilution, and clot formation was evaluated. A standard LAL gel clot assay with LPS (0–50 EU/ml) or E. coli (0–5.0 × 10⁵ CFU/ml) was also performed as a positive control. A clot remaining at the bottom of the tube after a gentle 180° inversion was considered positive, whereas if the sample remained liquid and flowed down the tube, the assay result was considered negative. Clots were rated as follows: firm, opaque gel remaining stable after 180° inversion (++); soft gel with moderate to considerate opacity and tube adhesion when rotated 90°; weak gel with slight-to-moderate opacity and adhesion of starch-like floccules to sides of the tube when slanted (±); and no visible increase in viscosity or opacity (–) (Bishop and White, 1986).

Individual HSC cohorts were used for each respective analysis. HSC Cohort A (n = 6) was used to measure differences in Oxy-Hc and amebocyte density over an 88-day period. HSC Cohort B (n = 6) was used to measure LAL activity in aquaculture HSCs (after 88 days of husbandry) in comparison to LAL from wild-type HSCs and a commercially available LAL kit (E-Toxate^TM, Sigma-Aldrich). HSC Cohort C (n = 4) was used to measure differences in amebocyte density, Hc concentrations, and LAL activity (after 6 months of husbandry) for the 10% bleed study. Cohort D (n = 4) served as a no-bleed control to the 10% bleed studies.

Individual HSC data (Oxy-Hc and amebocyte density) varied slightly from animal to animal but with a low standard deviation. The LAL activity derived from individual HSCs was also similar, with low standard deviation. For simplicity and assay accuracy, LAL activity was measured using pooled samples from the entire aquaculture cohort, composed of equal amounts of proteins from individual HSCs. The standard deviation reported for LAL activity represents each assay, performed in triplicate.

The RAS used for this HSC study was designed for conventional aquaculture operations and scalability using standard equipment. Specifically, it was configured to house 24 HSCs, with six in each of four groups. The RAS design, methods, and HSC viability with respect to repetitive bleeding and catheter tolerance were established during the study period (Figure 2).

Determining and maintaining optimal water quality parameters were especially important in assessing the feasibility of an indwelling catheter (Figure 3). Water temperatures were between 17.1 and 21.7°C (desired range: 15.0–25.0°C); pH was maintained from 7.5 to 8.0 (desired range: 7.5–8.8); and dissolved oxygen (DO) remained within 6.3–11.1 mg/L (desired range: >3.0–5.0) across all tanks. The target ammonia concentration was established at <1.0–3.0 ppm, with the average calculated at 0.732 (± 0.206) ppm across the study; and it stayed within the determined cautionary range (Timmons, 1994; Schreibman and Zarnoch, 2009; Ebeling and Timmons, 2010; Shelley and Lovatelli, 2011).

All HSCs in the RAS thrived throughout a 12-month period, as measured by biologic markers, including LAL activity, bleeding frequency, and breeding behaviors (i.e., copulation and sperm and egg release) and remained healthy throughout the investigation (100% survival). Water parameters did not change significantly throughout the course of the study at the calculated HSC density and feeding rate. Immediate and long-term action plans for system perturbations were established and deemed convenient and effective. The results indicated that HSC RAS husbandry is practicable, that management protocols can be easily executed using conventional equipment, and that available feed inputs are nutritionally robust and affordable (feed cost per pound: $0.78).

Intravascular catheters were inserted into the respective pericardial membranes of the HSCs, which allowed for routine hemolymph collection up to three times per month (see HSC Health Assessment After Repetitive Hemolymph Collection section; Figure 7) without repetitive membrane punctures (Figure 4). The procedure and device design were optimized over time and well tolerated by the HSCs, as all survived catheter implantation and continued to demonstrate robust health after surgery, as reflected in subsequent testing. The following parameters were analyzed at predetermined intervals (see HSC Health Assessment After Repetitive Hemolymph Collection section) throughout the study: total serum protein concentration; amebocyte concentration (cells/ml); and the LAL activity of amebocyte extracts in response to LPS.

Hemocyanin concentration was examined over the course of 88 days (HSC Cohort A; n = 6). The proportion of Hc with dioxygen-bound Hc (i.e., Oxy-Hc) is an indicator of health status, and it remained within a desirable range of 85–95% across all HSCs (Figure 5A). Amebocyte concentration was used to assess immune function and as an indication of well-being versus stress (Coates et al., 2012). The HSC amebocyte density levels remained high throughout the study and correlated with optimal vitality (Figure 5B; Day 0: 1.806E⁺⁰⁷; Day 88: 2.093E⁺⁰⁷).

The aquaculture LAL (HSC Cohort B; n = 6) had similar activity to that of the wild HSCs, and both extracts revealed slightly increased activity compared with the lyophilized and preserved LAL from the commercial kit (although values were not determined to be statistically significant; Figure 6A). The activity of aquaculture LAL remained constant throughout the 88-day culture period (data not shown). At higher concentrations (150 μg), the enzymatic activity of aquaculture LAL was twofold higher than that of the commercial LAL at an LPS concentration of 50 EU/ml (Figure 6B). These findings suggest that fresh aquaculture lysates may contain a greater abundance of clotting factors per microgram of total extract than that contained by commercial sources, as evidenced by clot formation after 1 h of incubation and as LPS concentrations become saturated, typical of a zero-order kinetic reaction.

Crude LAL derived from the aquaculture cohort (HSC Cohort B; n = 6) and the wild control demonstrated similar protein banding; however, different patterns were also observed with the commercial LAL using Coomassie-stained tris-glycine denaturing gel (Figure 6C). One protein band, identified as the L-chain of Factor C, migrated faster in the commercial LAL (∼43 kDa) than in the aquaculture LAL (∼52 kDa), which was confirmed via Western blotting with rabbit polyclonal antibody (COCH, ABclonal, Inc., Woburn, MA, United States; data not shown). Factor C typically generates a 43 kDa subunit (L-chain) under denaturing conditions (Shibata et al., 2018), and thus, the aquaculture-derived Factor C L-chain subunit could migrate at a higher molecular weight due to posttranslational modifications. As lysates prepared from freshly harvested wild HSCs revealed, this higher molecular weight form of Factor C does not appear to be a consequence of aquaculture but rather an effect of commercial LAL preparation (Figure 6C).

The effects of aquaculture on Hc and amebocyte rebound kinetics were evaluated after repetitive HSC harvests (10% of total hemolymph volume). Hemolymph was drawn at three intervals: Bleed 1 (10% hemolymph volume at time = 0); Bleed 2 (10% hemolymph volume at time = 16 days); and Bleed 3 (1 ml of hemolymph at time = 23 days, health assessment; Figure 7). HSC amebocyte density, total serum protein concentration, and LAL activity were thus measured to assess the impact of consecutive bleeding at t₀ = 0 days, t₁ = 16 days, and t₃ = 23 days (n = 4; Figure 7).

Average serum protein levels recovered after each successive 10% hemolymph extraction (t₀ vs. t₁ = 10.2% Δ; t₁ vs. t₂ = 6.7% Δ; Figure 7A). Amebocyte density decreased slightly (−10.7% Δ) after the first collection (t₀ vs. t₁) and rebounded above baseline after the second (12.8% increase between t₁ and t₂, however; this increase was associated with a higher standard deviation among individual HSCs). LAL activity was notably higher after the second instance over the 7-day recovery period (t₀ vs. t₁ = −14.9% Δ and t₁ vs. t₂ = + 5.0% Δ; Figure 7C). As a “no-bleed” control, a health assessment specimen (1.0 ml) from a second HSC cohort (n = 4) was taken at the same time and analyzed in parallel, with no significant change in hemolymph parameters (data not shown).

Aquaculture LAL was evaluated for detection of endotoxins (LPS-spiked) in chemically treated blood samples and compared to commercial standards. The previously described method (Bishop and White, 1986) to evaluate clot integrity was followed [Figure 8: firm gels (++); soft gels (+); weak clots (±); and no gelation (–)]. LPS-spiked blood samples and LAL commercial standards yielded comparable degrees of gelation from 0.5 to 50.0 EU/ml (Table 1). Strong clots (++) formed at endotoxin levels of 12.5–50 EU/ml; soft gel clots (+) formed at 3.125–6.25 EU/ml, and weak gel clots (±) formed at 0.5 EU/ml in both LPS-spiked blood samples and LAL commercial standards, while at the lowest concentration of 0.05 EU/ml, no gel clots were discernible [(–) clot rating; Table 1]. The neat (negative) blood samples remained viscous yet revealed no gelation (–).

Human blood with and without anticoagulant (heparin) demonstrated similar results: chemical treatment of the blood was required in order to detect LPS; whereas gel clots did not form in untreated blood samples (with or without anticoagulant) across the range of endotoxin concentrations (up to 50 EU/ml). This was consistent with previous studies showing that human blood inhibits LAL reactions (Martel et al., 1985; Lonza Walkersville Inc, 2014).

When E. coli was added to treated blood at concentrations between 0.5 and 500,000 CFU/mL, a firm gel clot formed [(++) clot rating; Table 2]. Interestingly, bacterial concentrations lower than 5,000 CFU/ml were not detectable in the absence of treated blood. To validate whether the treatment altered LPS activation of LAL, the LPS (0.05–50 EU/ml) was also added following chemical treatment, and clot formation was consistent with that of LPS spiked-blood samples prior to treatment (data not shown). Blood spiked with both free LPS and bacteria (exponential-phase) activated LAL clot formation (Tables 1, 2) with no false positives or false negatives observed. The “broth” control (no bacteria) did not form a gel clot, also suggesting that the bacterial growth medium was sufficiently washed away from the bacteria in the procedure and did not cause a false positive in the bacteria-spiked samples (Table 2).