11-Deoxycortisol controls hydromineral balance in the most basal osmoregulating vertebrate, sea lamprey (Petromyzon marinus)

It is unknown whether and how osmoregulation is controlled by corticosteroid signaling in the phylogenetically basal vertebrate group Agnatha, including lampreys and hagfishes. It is known that a truncated steroid biosynthetic pathway in lampreys produces two predominant circulating corticosteroids, 11-deoxycortisol (S) and 11-deoxycorticosterone (DOC). Furthermore, lampreys express only a single, ancestral corticosteroid receptor (CR). Whether S and/or DOC interact with the CR to control osmoregulation in lampreys is still unknown. We examined the role of the endogenous corticosteroids in vivo and ex vivo in sea lamprey (Petromyzon marinus) during the critical metamorphic period during which sea lamprey increase osmoregulatory capacity and acquire seawater (SW) tolerance. We demonstrate in vivo that increases in circulating [S] and gill CR abundance are associated with increases in osmoregulatory capacity during metamorphosis. We further show that in vivo and ex vivo treatment with S increases activity and expression of gill active ion transporters and improves SW tolerance, and that only S (and not DOC) has regulatory control over active ion transport in the gills. Lastly, we show that the lamprey CR expresses an ancestral, spironolactone-as-agonist structural motif and that spironolactone treatment in vivo increases osmoregulatory capacity. Together, these results demonstrate that S is an osmoregulatory hormone in lamprey and that receptor-mediated discriminative corticosteroid regulation of hydromineral balance is an evolutionarily basal trait among vertebrates.


Results
Lamprey gill CR has high affinity and specificity for 11-deoxycortisol over other endogenous and later-evolved corticosteroids. The physiological corticosteroids in lamprey, S DOC, are produced from a truncated steroid biosynthesis pathway in lieu of the existence of cyp11b genes in the lamprey genome (Fig. 1A). We performed classical receptor binding assays on larval and metamorphic lamprey gill tissue ex vivo to characterize binding properties of the lamprey gill CR. We also assessed binding affinity of the lamprey gill CR to endogenous (S and DOC) and later-evolved (F and A) corticosteroid ligands. Saturable binding by S in the lamprey gills produced curves used to calculate specific binding (B S ), dissociation constant (K d ), and abundance (B max ) for the lamprey gill CR (representative curves presented in Fig. 1B,C). The binding affinity hierarchy of the lamprey gill CR to corticosteroids was S > DOC > > F = A (Fig. 1D); the lamprey gill CR had significantly greater binding affinity (IC 50 ; half-maximal inhibition) for S (15.7 nM) than DOC (89.1 nM), or F and A (no binding detected).

11-Deoxycortisol treatment improves osmoregulatory capacity in metamorphosing lamprey.
In mid-metamorphic (Oct) lamprey, after the significant increase in gill CR but before the endogenous Scientific RepoRtS | (2020) 10 (Fig. 3B). Similar increases in plasma [Cl − ] (to ~ 155 mM) were observed in the Veh controls transferred to 25‰ (hereafter referred to as "SW") ( Fig. 4A). After 24 h exposure to SW, plasma [Cl − ] in lamprey treated with 10 and 50 μg g −1 S were both maintained closer to FW control levels (~ 140 and ~ 120 mM, respectively) and significantly lower than in the Veh control (Fig. 4A).  www.nature.com/scientificreports/  www.nature.com/scientificreports/ Lamprey administered S had significantly higher (twofold) gill NKA activity than the Veh control ( Fig. 4B), as well as significantly higher gill nka (Fig. 4C) and nkcc1 (Fig. 4D) mRNA abundance. Lamprey administered 50 μg g −1 S had increased gill NKA (1.5-fold; Fig. 4E) and NKCC1 (2.5-fold; Fig. 4F) protein abundance over the Veh control. Explanted gill tissue from mid-metamorphic lamprey incubated with S for 24 h ex vivo exhibited a dose-dependent increase in nka mRNA expression (Fig. 5A) and a similar trend for nkcc1 mRNA expression (Fig. 5B). Larval lamprey treated with either dose of S showed no significant increase in gill NKA activity (P = 0.297; one-way ANOVA) over the Veh control (Fig. S1). After 24 h exposure to 12‰, larvae plasma [Cl − ] was elevated from ~ 85 to 110 mM (P < 0.001) but there was no significant osmoregulatory improvement from prior treatment with S (P = 0.262).

11-Deoxycortisol controls osmoregulation in lamprey.
In a similar in vivo experiment using intraperitoneal administration of corticosteroids in metamorphic lamprey, we sought to test whether the endogenous corticosteroids in lamprey, S and DOC, controlled osmoregulation. As expected, intraperitoneal administration of S resulted in a dose-dependent elevation of circulating [S] (Fig. 6A). Compared to the Veh control (~ 2 μmol ADP mg −1 h −1 ), gill NKA activity was significantly elevated by treatment with S, but treatment with DOC did not elevate gill NKA activity (Fig. 6B). Similarly, gill NKA and NKCC1 protein abundance were significantly elevated (~ threefold and ~ sixfold, respectively) by treatment with S but neither were elevated by treatment with DOC ( Fig. 6C,D).
Ancestral MR-like structural properties of the lamprey CR confer osmoregulatory control by spironolactone. In a final experiment, we sought to relate structural properties of the lamprey CR to www.nature.com/scientificreports/ osmoregulatory function. It was recently determined that a key Leu-to-Thr (Ser in rodents) substitution in helix 8 of the vertebrate MR is critical for switching the action of spironolactone from an agonist to an antagonist. In a phylogeny of vertebrate corticosteroid receptors, the lamprey CR occupies a position that precedes the divergence of MR and GR (Fig. 7A). The lamprey CR possess the ancestral Leu at residue 810 (on human MR) on helix 8 (Fig. 7B), suggesting that spironolactone should be an agonist of the lamprey CR. Consistent with this prediction, in vivo treatment with spironolactone increased gill NKA activity (Fig. 7C).  www.nature.com/scientificreports/

Discussion
As the most phylogenetically basal vertebrates with an osmoregulatory strategy, lampreys are the ideal model organism to gain insight into the appearance and evolution of endocrine control of hydromineral balance in vertebrates. This work provides several novel lines of evidence that S is the primary corticosteroid hormone that acts through the ancestral CR to control osmoregulation in the basal vertebrate, sea lamprey: (i) plasma [S] and gill CR abundance are upregulated during metamorphosis when endogenous increases in salinity tolerance occur; (ii) plasma [S] is upregulated during SW exposure; (iii) ex vivo and in vivo treatment with S upregulates expression of critical active gill ion transporters and increases SW tolerance; (iv) in vivo treatment with the other known circulating corticosteroid in lamprey, DOC, does not upregulate such transcellular ionoregulatory machinery; and (v) treatment with a pharmacological agonist of the lamprey CR, spironolactone, is effective in stimulating lamprey gill ion transporter activity similar to the effects of S. To understand how corticosteroids control osmoregulation in lamprey, it is necessary to observe the nature of endogenous osmoregulatory and endocrine changes that occur naturally during metamorphosis and SW acclimation. In the present study, we show large increases in gill NKA activity during the later stages of metamorphosis (Oct and Nov) occurred along with increases in plasma [S] and gill CR abundance. The increase in gill NKA activity during metamorphosis corresponds with what has been observed previously with gill NKA (activity and protein abundance 8-10 ) as well as gill NKCC1 protein abundance 9 . An increase in gill CR abundance was detected earlier in metamorphosis (Oct) compared to that of plasma [S] (Nov). It makes intuitive sense that an increase in the intrinsic capacity of the lamprey gill to receive a corticosteroid signal (i.e., increase in gill CR abundance) would occur prior to an elevation of the concentration of circulating corticosteroid, whereas the opposite case (upregulating plasma [S] before increasing gill CR abundance) would less efficiently utilize the circulating hormone. A similar endocrine program is present in anadromous salmonids in which, along with circulating cortisol, GR abundance (as measured by binding studies similar to those used in the present study) is upregulated during the parr-smolt transformation ahead of SW entry [26][27][28][29][30] . In teleost fishes, the GR is known to be important in mediating osmoregulatory adaptation 23,31 . The hyperbolic relationship between lamprey gill NKA and plasma [S] shown in the present study suggests a saturating potential of the hormone-at some concentration of circulating S, the CRs in the gill become saturated and a further increase in plasma [S] does not further stimulate gill NKA activity. Indeed, the apparent saturating effect of [S] on gill NKA activity occurs www.nature.com/scientificreports/ around ~ 5-10 ng mL −1 (NKA activity at 50% = 10.7 ± 5.1 ng mL −1 ), or ~ 15-30 nM, which closely approximates the range of calculated values for IC 50 of the gill CR to S of ~ 15 nM. Conversely, gill NKA activity was more directly related to increases in gill CR abundance, suggesting that the receptor is important in directly mediating the corticosteroid signal and affecting an osmoregulatory response in the metamorphic sea lamprey. In anadromous salmonids, increases in GR abundance during the parr-smolt transformation is regulated by growth hormone (GH) 32 . GH has been identified in sea lamprey 33 , but more work is needed to determine whether osmoregulation is controlled by a similar endocrine program in sea lamprey. Mid-metamorphic (early-Oct) lamprey exposed to elevated salinity elicit a salinity-dependent increase in both plasma [Cl − ] and [S] (Fig. 3), whereas it is known that fully metamorphosed lamprey exhibit little to no loss of ion homeostasis after SW exposure 10 . The increase in plasma [S] upon SW exposure indicates that SW exposure early in metamorphosis (before hypo-osmoregulatory preparation is complete) is either stressful and/ or that circulating [S] increases to promote salt secretion. In either case, S signaling seems important for lamprey to initiate osmoregulatory changes necessary to survive in SW when osmoregulatory capacity is low.
Ex vivo and in vivo experimentation with cortisol have proven a useful method for establishing its role in osmoregulatory control in other vertebrates 14 , and so we used similar approaches to characterize the osmoregulatory role of S in sea lamprey. Our hormone treatment and salinity experiments were carried out in October, which we had determined to be the best time to conduct these experiments as it was after the significant increase in gill CR abundance, but prior to the endogenous increases in plasma [S] and peak gill NKA activity. During this time frame, SW tolerance is higher than in larvae (which can only tolerate a maximum of 12‰) but still lower than fully transformed juveniles 9 . After 24 h in elevated salinity, metamorphosing lamprey in October experienced salinity-dependent loss of osmotic homeostasis, evidenced by salinity-dependent increases in plasma [Cl − ] to well above ~ 120 mM [Cl − ] (Figs. 3, 4), which is the maximum ion disturbance experienced by fully metamorphosed juveniles 8 . Prior treatment with S clearly improved osmoregulatory capacity as evidenced by an enhanced ability to maintain lower levels of plasma [Cl − ] after SW exposure compared to Veh-injected controls. Differences observed between T 0 and Veh in gill ion transporter activity and expression reflects the upregulation of these osmoregulatory machinery during the natural progression of metamorphosis 8,9 and is not indicative of an artifactual effect of the Veh. The effect of S treatment on plasma [Cl − ] after SW transfer was dose-dependent. If S signaling for gill osmoregulatory changes is indeed CR-mediated, this dose-dependency indicates that gill CR is present at levels sufficient to respond to a range of circulating S concentrations. The enhanced osmoregulatory ability in lamprey treated with S is likely due to the corresponding increases in gill ion transporters (NKA and NKCC1) observed, since these transporters also increase during metamorphosis when salinity tolerance increases. Similar increases in gill NKA and NKCC1 have also been observed in conjunction with higher levels of salinity tolerance in teleost fishes treated with cortisol 15 .
In addition to characterizing the role of S and CR during metamorphosis, we sought to investigate the specificity of corticosteroid action in vivo by conducting a similar experiment administering both endogenous corticosteroids, S or DOC. We confirmed that our method of intraperitoneal injection with S achieved dose-wise increases in circulating [S] within a physiological range, and replicated the results from Fig. 4 showing an increase in gill NKA and NKCC1 occurs after treatment with S. In contrast to the stimulating effects of S, we observed that the other endogenous corticosteroid, DOC, did not significantly upregulate such transcellular active gill ionoregulatory machinery. The lack of effect of DOC on active gill ion transport is likely due to the relatively low binding affinity for DOC compared to S by the lamprey gill CR, as first reported by Close et al. 1 and replicated in the present study. It is also possible that in situ inactivation of DOC by a reductase (e.g., 5α-reductase) or some other pathway of steroid catabolism is occurring in the gill, much like how hydroxysteroid dehydrogenases (HSDs) are known to regulate tissue-specific physiologic actions to corticosteroids in other vertebrates 34 .
It must be acknowledged that, in addition to its role in transcellular active ion transport demonstrated in the present study, S may also have a role in lamprey osmoregulation through other mechanisms such as control of paracellular ion and water transport. It is known that tight junction (TJ) complexes dynamically control paracellular permeability in many vertebrate epithelia, including changes in gill epithelium of teleost fishes from "tight" in FW to "leaky" in SW 35 . A functional role for TJ complex proteins has more recently been elucidated in the sea lamprey [35][36][37][38] . Several studies have demonstrated putative corticosteroid control of TJ proteins in cultured teleost fish gill epithelia 24,39,40 , and it is possible that S may thus also control TJ complexing and paracellular permeability in the sea lamprey gill, in addition to its role in controlling active ion transport.
Interestingly, we were unable to stimulate an osmoregulatory response in larval lamprey with an identical in vivo approach of S treatment, despite larvae having apparently adequate gill CR abundance (B max ~ 100-150 fmol mg −1 ) to bind the administered S. It could be that cell type-specific expression of the lamprey CR is modulating differences in osmoregulatory action of S between larvae and metamorphic lamprey-that the CR is expressed in Cl − secreting gill ionocytes in metamorphic lamprey but not in larval lamprey. Immunological studies using antibodies raised against the lamprey CR are needed to better understand the physiological role and localization of the lamprey CR. Another explanation may be that signaling from a GH/insulin-like growth factor (IGF) pathway is required for initial ionocyte differentiation and upregulation of CR early in lamprey metamorphosis, after which the corticosteroid can then act to upregulate protein abundance of ion transporters, as has been a suggested model of coordinated endocrine control of osmoregulation in teleosts 41 . In any case, the role of S in controlling osmoregulation in metamorphic lamprey is clearly apparent-the correspondence of plasma [S] and calculated gill CR K d and IC 50 values, the concordant rise in systemic S and gill CR, and the specific S effects on the gill ionoregulatory apparatus combine to suggest S is an important hormone contributing to development of SW tolerance in the basal sea lamprey.
Lastly, we sought to obtain further in vivo evidence demonstrating that the CR is important in mediating the osmoregulatory control of S during lamprey metamorphosis. In mammalian research, antagonists of MR (such as spironolactone and eplerenone, derivatives of progesterone) or GR (such as RU486) have been used for www.nature.com/scientificreports/ more than half a century, due to their anti-aldosterone, -cortisol, and -progesterone activities 42,43 . In fishes, MR and GR antagonists have been used to examine the osmoregulatory roles of the MR-and GR-mediated action of corticosteroids, namely cortisol, with mixed results 18,27,31,[43][44][45][46][47][48] . Recently it has been shown that spironolactone serves as an agonist, not an antagonist, to the MR in basal lineages of fishes 49,50 . It was discovered that a substitution of Leu by Thr (Ser in rodents) on helix 8 of the MR of tetrapods confers the switch of progesterone and its derivative, spironolactone, from MR agonists to antagonists, and it has been suggested that this molecular switch may be causally related to the appearance of aldosterone synthesis and mineralocorticoid function of aldosterone in tetrapods 51 . Analysis of the lamprey CR revealed the key Leu residue was present on helix 8, and thus we should expect that spironolactone would act as a CR agonist. If the osmoregulatory effects of S treatment were indeed acting through the lamprey CR, we would expect to see a similar effect of spironolactone. We observed this expected rise in gill NKA activity after 3 weeks of in vivo treatment with spironolactone (Fig. 6), similar in magnitude to our in vivo experiments with S (Fig. 4). These results suggest an MR-like functional quality of the lamprey CR, which has been previously suggested based on molecular and structural analyses of the lamprey CR 2,25,52,53 . To our knowledge, our in vivo results demonstrating the stimulatory action of progesterone derivatives on osmoregulation are the only such in vivo evidence in any vertebrate lineage predating the actinopterygian-sarcopterygian split.
In conclusion, the present study provides much needed in vivo evidence and physiological context to earlier investigations and conflicting reports regarding corticosteroid function in the basal Agnathans. Our work underscores the importance of in vivo and ex vivo studies and inspires more research into corticosteroid function in Agnathan physiology. We demonstrate that a receptor-mediated signal from S, and not DOC, is a major endocrine process controlling osmoregulatory processes in the sea lamprey. That lamprey exhibit discriminative corticosteroid control of osmoregulation (i.e., stimulation active ion transport processes by S but not DOC) indicates that such corticosteroid function appeared early in vertebrate evolution, perhaps in association with the appearance of an osmoregulatory strategy, which has been maintained in nearly all later-evolved vertebrate lineages.

Materials and methods
Experimental subjects and animal care. All animal care and experimentation were carried out using protocols approved by the Internal Animal Care and Use Committee at the University of Massachusetts and the U.S. Geological Survey (Protocol number: 2016-0009) and in accordance with all relevant guidelines and regulation (Guide for Care and Use of Laboratory Animals, National Research Council, Washington, DC). Wild larval and pre-metamorphic lamprey were caught in July by electrofishing on the Sawmill River, a tributary to the Connecticut River in Western Massachusetts, USA. After capture, animals were held in 1.5 m diameter flowthrough tanks supplied with Connecticut River water with 10 cm of sand. For the metamorphic profile, larval and metamorphosing lamprey were randomly selected for sampling each month from August to December. For hormone treatment experiments, mid-metamorphic lamprey were held in aquaria (80 L) containing dechlorinated, filtered, and aerated recirculating municipal fresh water kept at 15 °C. Artificial SW dilutions were made using a commercial sea salt mix (Crystal Sea Salt, USA).
In vivo and ex vivo experimentation. Salinity tolerance testing and in vivo hormone treatments were performed in early October on mid-metamorphic (early stage 7) 6 . The biometric data of the lamprey in these experiments were: length = 14.7 ± 0.3 cm; body mass = 5.4 ± 0.4 g; sex ratio = 50% male, 50% female. In salinity tolerance tests, lamprey were exposed to varying salinities (15, 20, or 25‰) or a FW control and sampled for blood after 24 h. For hormone treatments, lamprey were anesthetized with MS-222 (100 mg L −1 buffered by NaHCO 3 , pH 7.4) and injected intraperitoneally with vehicle alone (Veh; 1:1, oil:shortening) or Veh containing S, DOC, or spironolactone, then held in freshwater for 12 days (21 days for spironolactone) prior to sampling in FW or after a 24 h exposure to 25‰ (referred to as "SW"). To account for changes which may be occurring naturally during metamorphosis, uninjected lamprey were sampled on the day of injections as a time = 0 control (T 0 ). The doses for S injection were adjusted slightly between experiments (Figs. 4 and 6) to better capture any potential dose-dependent effects of S on gill NKA activity. Lamprey were sampled 21 days after spironolactone injection was to allow additional metamorphic progress to occur between injection and sampling in order to better observe the possibility of spironolactone acting as an antagonist of the lamprey CR and slowing osmoregulatory changes during metamorphosis.
For ex vivo hormone treatment, intact whole gill pouches were excised from FW-acclimated mid-metamorphic lamprey and carefully dissected to unfold them from their basket-like arrangement and cleanly remove the afferent and efferent ends to expose the internal filamental vasculature to incubation media. The final ex vivo branchial unit was 4-8 gill filaments (~ 1 mg of wet tissue). Groups of 6-10 intact filament pieces were placed in incubation media (DMEM containing 5 mM glucose and 100 U mL −1 penstrep) containing dissolved hormone such that gill tissue from a single fish would be exposed to every hormone treatment (0.05-5.0 μg mL −1 S) and a media-only control (Ctrl). Doses of S for ex vivo culture were chosen according to doses of cortisol used in similar ex vivo organ culture experiments in teleosts 54  www.nature.com/scientificreports/ were placed in SEI buffer (150 mM sucrose, 10 mM EDTA, 50 mM imidazole, pH 7.3) before freezing. Plasma [Cl − ] was measured using a digital chloridometer (Haake Buchler Instruments Inc., USA).
Na + /K + -ATPase activity assay. NKA activity was determined by measuring the ouabain-sensitive ADP production of gill homogenates 55 . Gill tissue was thawed, homogenized in SEID buffer (SEI with 0.1% deoxycholate) and centrifuged at 3,000g for 5 min. The resulting supernatant was used in an enzyme-linked kinetic assay, which couples ADP production to NADH reduction in a 1:1 ratio to determine ATPase activity. Protein concentration was determined spectrophotometrically using a bovine serum albumin (BSA) standard curve (BCA Protein Assay, Pierce, USA) and the ouabain-sensitive ATPase activity expressed as μmol ADP mg protein −1 h −1 . Radioreceptor binding assays for analysis of lamprey gill CR. Corticosteroid receptor binding assay was modified from previous methods 1, 21 . Frozen gill tissue pooled from two individuals (~ 8-10 gill pouches) were placed in 300 μL of ice-cold HEPES assay buffer (25 mM HEPES. 10 mM NaCl, 1 mM monothioglycerol, pH 7.4) and homogenized on ice using a ground glass homogenizer. Homogenates were centrifuged for 10 min at 2,000g at 4 °C and the resulting supernatant (crude cytosolic preparation) was placed on ice for use in receptor binding assay which was carried out on ice in a 96-well plate. An aliquot of supernatant was reserved to determine protein concentration (BCA Protein Assay, Pierce). For each sample, total (B T ) and non-specific Gill mRNA and protein analyses. Gill NKA and NKCC1 mRNA and protein abundance were analyzed by standard quantitative real-time polymerase chain reaction (qPCR) and Western blotting using procedures, molecular primers, and antibodies as previously validated and described in our laboratory 9  To understand the nature of the relationships between hormone or receptor and the developmentally regulated rise in gill NKA activity, we compared three theoretical curve-fitting analyses (linear, hyperbolic, and quadratic) to determine the simplest model for the relationship (most degrees of freedom) that best fit (highest r 2 ) the data. Normality and homogeneity of variance assumptions were tested using Shapiro-Wilk and Levene's tests, respectively. Treatment effects and comparisons between treatment groups were analyzed using Student's t-test,