Stony Coral Tissue Loss Disease in Florida Is Associated With Disruption of Host–Zooxanthellae Physiology

Samples from eight species of corals (Colpophyllia natans, Dendrogyra cylindrus, Diploria labyrinthiformis, Meandrina meandrites, Montastraea cavernosa, Orbicella faveolata, Pseudodiploria strigosa, and Siderastrea siderea) that exhibited gross clinical signs of acute, subacute, or chronic tissue loss attributed to stony coral tissue loss disease (SCTLD) were collected from the Florida Reef Tract during 2016–2018 and examined histopathologically. The hallmark microscopic lesion seen in all eight species was focal to multifocal lytic necrosis (LN) originating in the gastrodermis of the basal body wall (BBW) and extending to the calicodermis, with more advanced lesions involving the surface body wall. This was accompanied by other degenerative changes in host cells such as mucocyte hypertrophy, degradation and fragmentation of gastrodermal architecture, and disintegration of the mesoglea. Zooxanthellae manifested various changes including necrosis (cytoplasmic hypereosinophilia, pyknosis); peripheral nuclear chromatin condensation; cytoplasmic vacuolation accompanied by deformation, swelling, or atrophy; swollen accumulation bodies; prominent pyrenoids; and degraded chloroplasts. Polyhedral intracytoplasmic eosinophilic periodic acid–Schiff-positive crystalline inclusion bodies (∼1–10 μm in length) were seen only in M. cavernosa and P. strigosa BBW gastrodermis in or adjacent to active lesions and some unaffected areas (without surface lesions) of diseased colonies. Coccoidlike or coccobacilloidlike structures (Gram-neutral) reminiscent of microorganisms were occasionally associated with LN lesions or seen in apparently healthy tissue of diseased colonies along with various parasites and other bacteria all considered likely secondary colonizers. Of the 82 samples showing gross lesions of SCTLD, 71 (87%) were confirmed histologically to have LN. Collectively, pathology indicates that SCTLD is the result of a disruption of host–symbiont physiology with lesions originating in the BBW leading to detachment and sloughing of tissues from the skeleton. Future investigations could focus on identifying the cause and pathogenesis of this process.

Samples from eight species of corals (Colpophyllia natans, Dendrogyra cylindrus, Diploria labyrinthiformis, Meandrina meandrites, Montastraea cavernosa, Orbicella faveolata, Pseudodiploria strigosa, and Siderastrea siderea) that exhibited gross clinical signs of acute, subacute, or chronic tissue loss attributed to stony coral tissue loss disease (SCTLD) were collected from the Florida Reef Tract during 2016-2018 and examined histopathologically. The hallmark microscopic lesion seen in all eight species was focal to multifocal lytic necrosis (LN) originating in the gastrodermis of the basal body wall (BBW) and extending to the calicodermis, with more advanced lesions involving the surface body wall. This was accompanied by other degenerative changes in host cells such as mucocyte hypertrophy, degradation and fragmentation of gastrodermal architecture, and disintegration of the mesoglea. Zooxanthellae manifested various changes including necrosis (cytoplasmic hypereosinophilia, pyknosis); peripheral nuclear chromatin condensation; cytoplasmic vacuolation accompanied by deformation, swelling, or atrophy; swollen accumulation bodies; prominent pyrenoids; and degraded chloroplasts. Polyhedral intracytoplasmic eosinophilic periodic acid-Schiff-positive crystalline inclusion bodies (∼1-10 µm in length) were seen only in M. cavernosa and P. strigosa BBW gastrodermis in or adjacent to active lesions and some unaffected areas (without surface lesions) of diseased colonies. Coccoidlike or coccobacilloidlike structures (Gram-neutral) reminiscent of microorganisms were occasionally associated with LN lesions or seen in apparently healthy tissue of diseased colonies along with various parasites and other bacteria all considered likely secondary colonizers. Of the 82 samples showing gross lesions of SCTLD, 71 (87%) were confirmed histologically to have LN. Collectively, pathology indicates that SCTLD is the result of a disruption of host-symbiont physiology with lesions originating in the BBW leading to detachment and sloughing of tissues from the skeleton. Future investigations could focus on identifying the cause and pathogenesis of this process.

INTRODUCTION
The Florida Reef Tract spans some 577 km from the Dry Tortugas (the southernmost Florida Keys) to Martin County in the north. With a diverse assemblage of more than 6,000 marine species, the Florida Reef Tract attracts more than 30 million visitors annually, with fishing, diving, and boating expenditures generating more than $6.3 billion to the State. 1 This ecosystem, however, has suffered significant losses of coral and associated organisms, starting with an unexplained die-off in Panama of the dominant grazing long-spined sea urchin Diadema antillarum in which more than 90% of the population throughout the Caribbean was wiped out within a year (1983)(1984), and populations have yet to recover (Lessios, 1988(Lessios, , 2016. In subsequent decades, a majority of the dominant reef coral Acropora (A. palmata, A. cervicornis) declined by about 80%, mainly due to tissue loss diseases (TLDs), leading to marked reductions in the three-dimensional complexity and biodiversity of Florida's coral reefs (Aronson and Precht, 2001). Since then, this region has experienced a plethora of coral diseases Porter et al., 2001).
Global climate change is likely to aggravate this situation. Warming temperatures induce corals to expel their dinoflagellate symbionts (zooxanthellae), a process known as bleaching (Coles and Jokiel, 1977). Since 1987, six mass coral-bleaching events have affected the entire Florida Reef Tract, with moderate intermittent incidents occurring annually since 2006 (except for 2013 when bleaching reports were low) (Manzello, 2015;Gintert et al., 2018). The Florida Reef Tract has also experienced an increasing prevalence of acute or chronic TLDs such as blackband disease, white-band disease, acroporid serratiosis, dark spot syndrome, and white plague (WP) (Kuta and Richardson, 1996;Richardson et al., 1998a,b;Patterson et al., 2002;Peters, 2015;Bruckner, 2016;Gignoux-Wolfsohn et al., 2020). The demise of acroporids, continuing intermittent chronic disease outbreaks, and coral loss due to bleaching have added to the concern over the stability and health of Florida's coral reef ecosystem (Williams and Miller, 2012) and has highlighted the previously underestimated role of disease (Miller et al., 2014).
In 2014, SCTLD appeared to spread following bleaching and sedimentation events (Manzello, 2015;Miller et al., 2016), and, despite much research, the etiology of SCTLD remains unknown. Along with the diversity of species affected, SCTLD can have different rates of tissue loss, differences in lesion morphology, and differences in the distribution of lesions in a colony . Subacute and chronic tissue-loss lesions can be observed with bleached borders, while acute lesions can manifest with normally pigmented tissue adjacent to denuded skeleton . All types of tissue loss are progressive, differing with species and region, and, in some instances, SCTLD appears transmissible, indicating an infectious etiology .
No systematic descriptions of SCTLD exist at the gross or the microscopic level. Such information might shed light on mechanisms or causes of disease, particularly in poorly understood animals such as corals (Work and Meteyer, 2014). Moreover, inferring the etiology of coral disease based solely on gross lesions might be misleading (Work and Aeby, 2006), particularly because TLDs might have multiple manifestations at the cellular level (Work et al., 2012). Our objective here was to systematically describe pathology of SCTLD at the gross and microscopic levels in eight coral species sampled during 2016-2018.

Study Sites
Tissue samples were collected from eight coral species (Colpophyllia natans, Dendrogyra cylindrus, Diploria labyrinthiformis, Meandrina meandrites, Montastraea cavernosa, Orbicella faveolata, Pseudodiploria strigosa, and Siderastrea siderea) from 13 sites at which SCTLD was epidemic or endemic (Rosales et al., 2020), ranging from Broward County southward along the Florida Reef Tract to the Lower Florida Keys. Reference samples from apparently healthy corals in regions without SCTLD at the time were collected at two timepoints: in April 2017 from the northernmost portion of the Florida Reef Tract from a site in Martin County and two sites in the Middle Keys, and in May-June 2018 from three sites in the Lower Keys, offshore of Key West. Sites spanned southeast Florida, offshore patch and fore reefs, and midchannel patch reefs in the Florida Keys (Figure 1). In 2017 and 2018, unaffected reference sample sites were distant from the SCTLD epidemic boundary but have since been affected.

Coral Sample Collection
Collection methods followed established protocols (Woodley et al., 2008). Samplers measured colony dimensions and recorded whether colonies were diseased or apparently healthy (i.e., FIGURE 1 | Sites along the Florida Reef Tract at which corals were sampled for stony coral tissue loss disease (SCTLD) histological analysis. Sites are in four regions (Southeast Florida [SE FL] and the Upper, Middle, and Lower Keys) and three general reef habitats (SE FL, offshore, and mid-channel patch). Project site codes = A-H used in Table 1, sampling dates = MM/YY, * = site sampled before the arrival of SCTLD.
those showing no evidence of tissue loss or discoloration) before sampling and photographing the whole colony and areas most suitable for a biopsy. A biopsy consisted of a circular core of tissue and skeleton collected using a 25.4mm-diameter sterilized stainless-steel corer-punch hammered into a colony until it reached a depth sufficient to obtain an ample sample, which varied with species. Individual reference sample tissue cores were collected from apparently healthy colonies (HH [histology healthy]) with no evidence of tissue loss or discoloration either at reference sites without SCTLD or from sites with SCTLD. For each diseased colony, one tissue core was collected from an apparently unaffected tissue area (HU [histology unaffected]), and a second core was taken from the margin between intact tissue and bare skeleton (HD [histology diseased]) ( Table 1). Apparently healthy corals were always sampled before diseased corals at sites with SCTLD, and a new pair of disposable nitrile gloves was worn for sampling each colony. After collection at each site, a photograph was taken of each biopsy core sample and postbiopsy area on the colony, and each core was placed into its own labeled Whirl-Pak sample bag and taken to the boat for further processing.
On board, sterile bone cutters were used to clip off a small representative portion of each core to be fixed for transmission electron microscopy (TEM). Samples were fixed for histology or TEM immediately after the dive, conditions permitting, or stored in a cooler filled with seawater at ambient temperature (and held in the shade) for transport to the laboratory (and then fixed within < 6 h from collection). To minimize cross contamination between sites during sampling, nitrile gloves were discarded, and all collection equipment sterilized on board in a 5-10% solution of sodium hypochlorite for 20 min.

Histopathology
Before histopathology, samples were photographed close up using a digital camera fitted with a macro lens (Nikon, Tokyo). Samples with grossly visible lesions were further examined under higher magnification with a dissecting microscope attached to an Olympus DP71 digital camera (Tokyo) and photomicrographs were taken. For ease of orientation and integrity, and to minimize loss of any surface microorganisms after the skeleton had been removed following decalcification, all HD samples were enrobed with 1.5% agarose in a heated vacuum oven (60 • C) at 22 mm Hg (Jones and Calabresi, 2007) (HU and HH samples were not enrobed with agarose but otherwise were processed as for HD samples), followed by decalcification with 10% ethylenediaminetetraacetic acid (EDTA-Na 2 ·2H 2 O, MW = 372.1; Fisher Scientific, Waltham, Pennsylvania) adjusted to pH 6.8-7.2 with NaOH (approximately 9 g per L of solution) on a shaker table (Price and Peters, 2018). Tissue enrobed in agarose was cut with a razor blade to expose the aboral (bottom) side facing the skeletal tissue opening, allowing contact with the EDTA solution. The EDTA solution was changed every 48-72 h until no resistance was felt in placing a razor blade into the tissue. Decalcification took 14-60 days depending on the size (depth of the skeleton) of each sample and longer if the tissue was enrobed. Decalcified tissue samples were then processed for routine histology.

Transmission Electron Microscopy
After histological evaluation, samples showing specific areas of interest were chosen for TEM study. Samples were decalcified with 10% EDTA solution for 8 days without agarose enrobing, cut into an appropriate size (approximately 1 × 1 × 1 mm), followed by postfixation with 1% osmium tetroxide in 1.25% NaHCO 3 buffer, pH 7.2, for 1 h at room temperature. Tissue samples were washed in Nanopure water three times for 5 min each and then dehydrated in a graded series of ethanol and embedded in Spurr epoxy resin (Sigma-Aldrich, St. Louis, Missouri). The epoxy block was sectioned for semithin sections with a Leica EM UC6 ultramicrotome (Leica Microsystems, Vienna) attached with a glass knife, at 1-µm thickness, adhered onto a glass slide, stained with toluidine blue, mounted with a cover glass, and examined under light microscopy for preparation quality and tissue condition. Ultrathin sections (90 nm) were cut with a diamond knife on the ultramicrotome and placed on copper grids. The sections were then stained with 2% uranyl acetate in 50% ethanol and lead citrate (Reynolds, 1963) and examined with a Jeol JEM-1400 transmission electron microscope (JEOL, Tokyo) equipped with a digital photomicrograph ORIUS TM SC1000 CCD camera (GATAN Inc., Pleasanton, California) in TEM brightfield mode at 80 kV.

RESULTS
A total of 263 samples, comprising eight species of corals from 17 sites (13 diseased and 6 reference sites [of which 2 were diseased the following year, sites A and E, and were re-sampled as diseased sites]) (Figure 1), were processed from 88 diseased colonies comprising 88 tissue-loss lesions (HD) and 82 grossly unaffected tissues (HU) from the same diseased colonies (except for Dendrogyra cylindrus which had no HU collected). An additional 42 unaffected tissues from apparently healthy (HH) colonies were collected in the same SCTLD-affected sites. Finally, unaffected tissues from 51 apparently healthy (HH) reference colonies were taken in SCTLD-free sites ( Table 1).

Gross Description
Grossly, SCTLD comprised varying shapes and sizes of amorphous distinct areas of tissue loss revealing (1) wide areas (∼50 mm or more width) of recent necrosis indicated by intact Frontiers in Marine Science | www.frontiersin.org FIGURE 3 | Gross acute to chronic lesions and discoloration in stony coral tissue loss disease. (A) Pseudodiploria strigosa with multifocal acute to subacute tissue loss. Note exposed skeleton bereft of tissues partly overgrown by green turf algae (black arrow) with area further from the tissue loss margin (white arrows) covered with particulate matter (arrowheads). (B) Montastraea cavernosa with subacute to chronic tissue loss. Note distinct to indistinct band of bare white skeleton behind active disease margin (black arrows) and followed by turf overgrowth (white arrows). (C) Orbicella faveolata with chronic tissue loss. Note area of tissue loss revealing bare white skeleton (black arrows) and what appears to be a white biofilm at the tissue-loss interface (white arrows), both followed by turf overgrowth (arrowheads). (D) Siderastrea siderea with multifocal to diffuse subacute to chronic tissue loss and pinkish discoloration, with outer edges of chronic lesions covered with mucus strands (arrows). Note tissue-loss areas (arrowheads) radiating from the polyps' oral regions and surrounding septa between adjacent polyps.  (3) narrow (∼10 mm or less width) to almost unnoticeable areas of recent necrosis apposed to skeleton overgrown with turf algae (chronic tissue loss, Figure 3C). Stony coral tissue loss disease in S. siderea appeared to present differently, with multifocal pinkish discoloration around the polyps and the surrounding septa, frequently with mucus strands present on the outer areas of the lesion ( Figure 3D). Location of tissue loss did not have any pattern; lesions appeared near the bottom, top, or edges of colonies regardless of species affected.
At the macroscopic level, the lesions differed in appearance depending on the anatomy of the affected species. In general, however, there was a distinct margin between healthy tissue and bare skeleton. Brain corals (C. natans, Diploria labyrinthiformis, and P. strigosa, family Faviidae) showed lesion margins that could traverse adjacent septa ( Figure 4A). In M. cavernosa and O. faveolata (families Montastraeidae and Merulinidae, respectively), tissue loss appeared to progress along the top of the septa, where the surface tissue is thin, allowing spreading across neighboring septa around the polyp while progressing across to a different polyp ( Figure 4B). In pillar or maze corals (Dendrogyra cylindrus and Meandrina meandrites, family Meandrinidae), borders of areas of tissue loss were more undulating ( Figure 4C). In S. siderea, pinkish discoloration associated with tissue loss appeared around the oral region and surrounding septa ( Figure 4D).

Histopathological and Ultrastructural Description
The low-magnification overview showed that lytic necrosis (LN) 6 first affected the basal body wall (BBW) gastrodermis ( Figure 5A) and progressed toward the surface body wall (SBW) (polyps and coenenchyme), presenting grossly as tissue loss. In advanced lesions, in P. strigosa, for example, necrosis appeared to advance across the full thickness of the SBW and BBW, forming a transitional boundary between degraded areas and intact tissue (Figure 5B).
At the cellular level, putative early lesions of SCTLD appeared to start with zooxanthellae in the otherwise intact BBW gastrodermis. In some apparently healthy corals from healthy reference (Figures 6A,B,D) or SCTLD-affected ( Figure 6C) sites, zooxanthellae were necrotic or swollen, with intracytoplasmic vacuolation of the gastrodermis, mesogleal edema, and vacuolation of the calicodermis (Figure 6A), with exocytosis of zooxanthellae ( Figure 6B), and pyknosis of the calicodermis (Figures 6B,C). In Montastraea cavernosa, individual or clusters of crystalline inclusion bodies (CIBs; details described below) were seen adjacent to zooxanthellae or vacuolated areas in the gastrodermis (Figures 6A-C). At the tissue level, putative early-stage lesions of SCTLD appear to start in the BBW of the septum or coenenchyme (Figures 6D-F), with early foci of initiating LN seen along the gastrodermis of the BBW (Figure 6D), advancing with a cleft forming between the BBW gastrodermis and the calicodermis ( Figure 6E) and progressing to multifocal BBW LN with necrotic tissue sloughing into the gastrovascular canal (GVC) (Figure 6F).
Presumptively in parallel with pathology in BBW zooxanthellae, cellular changes in the SBW were seen, with zooxanthellae initially affected, followed by SBW gastrodermal pathology. Zooxanthellae exhibited necrosis with dilation of symbiosomes, and gastrodermis had intracytoplasmic vacuolation ( Figure 7A), with zooxanthellae showing accumulation of intracytoplasmic refractile granules, hypopigmentation, deformation, or fragmentation (Figures 7B-F). The lesion seemed to progress to enlargement of the symbiosome surrounding zooxanthellae (Figures 7C-D), some of which manifested coagulation necrosis of individual zooxanthellae ( Figure 7C) and coalescence of symbiosomes ( Figures 7A,D-F). In advanced lesions, there was diffuse LN of the gastrodermis of the SBW and BBW ( Figure 7E), with necrotic zooxanthellae that occasionally stained positive (presumptive melanin, dark brown to black) with FM (Figures 7E,F).
On TEM in M. cavernosa, degraded coral gastrodermal cells were seen, with vacuolated cytoplasm filled with debris, phagosomes, degenerating nuclei, shrunken zooxanthellae with a loss of symbiosome membranes, and a general collapse of the symbiont-host cell integrity ( Figure 8A). Mucocytes were in somewhat better condition ( Figure 8A). Zooxanthellae showed varying degrees of nuclear fragmentation and degeneration with the formation of condensed peripheral chromatin (Figures 8B-D), degrading chloroplasts with loss of thylakoid structure (Figure 8C), and enlarged accumulation bodies with cellular debris (Figure 8D).
Advancement of LN into the SBW was observed initially in the gastrodermis. Distention of the symbiosome surrounding the zooxanthellae expanded beyond the basement membrane of the gastrodermis and into the mesoglea, leading to its fragmentation (Figures 9A,B). This progressed to full-scale gastrodermal necrosis, with release of cell debris and zooxanthellae into the GVCs (Figures 9C,D), mesogleal vacuolation and necrosis (Figure 9), and progression of vacuolation beyond the epidermal basement membrane (Figure 9C). Lesion advancement was followed by collapse of the mesoglea and sloughing of the epidermis (Figure 9D).
In Meandrina meandrites, Montastraea cavernosa, P. strigosa, and O. faveolata, initiation of necrosis of the gastrodermis of the SBW often occurred near areas of advanced necrosis of the     BBW (Figure 10), supporting the hypothesis that lesions start in the BBW and progress to the SBW. Advanced LN lesions in the BBW were associated with initial SBW gastrodermal necrosis (Figures 10A-C) and tissue sloughing into the GVC (Figure 10C), to mesogleal disruption and gastrodermal atrophy ( Figure 10D).
End-stage lesions at the tissue level involved SBW epidermal necrosis and sloughing (Figure 11A), full thickness necrosis of all tissue layers of areas of the BBW and the SBW with segmental ( Figure 11B) to diffuse (Figure 11C) distribution, necrosis of the mesenterial filaments (Figures 11D,E), and necrosis of the tentacular gastrodermis ( Figure 11F). In S. siderea, mesenterial calicoblasts with eosinophilic granules (coral acid-rich proteins, CARPs) were prominent (Figure 11B, inset).
Other less consistent findings were seen on histology. Intracytoplasmic polyhedral CIBs (∼1-10 µm in length) were seen mostly in the BBW and, occasionally, SBW gastrodermis of apparently healthy and diseased M. cavernosa and P. strigosa. These stained positive with PAS (Figures 12A,B) and were sometimes associated with LN ( Figure 12B) or, in M. cavernosa, presumptive early-stage lesions (Figure 6A-C). Crystalline inclusion bodies stained dark green with toluidine blue ( Figure 12C) and Feulgen, pale blue with thionin, and black with FM, but negative with GMS, Prussian blue, Macchiavello, distinctly delineated from intact BBW to either side (boundary, gray arrowheads). Note also LN of adjacent gastrodermis of nearby SBW (black arrow) and sloughing of necrotic tissue and zooxanthellae into the GVC (double-line arrow). (D) BBW LN in Orbicella faveolata. Note same lesion as in (C) for BBW (arrowhead), but, in contrast, see localized attenuation and atrophy of nearby gastrodermis of SBW (arrow). Hematoxylin and eosin (H and E). bbw, basal body wall; c, calicodermis; e, epidermis; g, gastrodermis; gvc, gastrovascular canal; m, mesoglea; sbw, surface body wall.
Other structures seen on histology in all species examined included clusters of either coccoidlike [dark blue to mauve in thionin (Figures 13A,D) or Giemsa (Figures 13B,C)] or coccobacilloidlike (pale to medium magenta in Giemsa, Figures 13B,C). Gram-neutral structures often distributed across multiple cells mainly in the gastrodermis of the BBW or SBW but occasionally in LN lesions ( Figure 13D). Various organisms were seen in healthy and diseased corals but were considered incidental findings because of their consistent lack of association with cell pathology. These included trichodinid ciliates on diseased and healthy O. faveolata, S. siderea, C. natans, P. strigosa, and D. cylindrus (Figure 13E    species. Endolithic organisms were occasionally observed penetrating or in close contact with the BBW of the coral tissue with advancing LN.

Prevalence of LN Lesions Across Species
Hallmark LN lesions were observed in samples collected from diseased (HD) and, to a lesser extent, unaffected areas on diseased colonies (HU), as well as in samples taken from apparently healthy colonies (HH) in disease-affected sites ( Table 1). Of the samples showing diverse gross signs of acute, subacute, or chronic tissue loss (HD) evaluated across species (n = 82), a prevalence of 87% was confirmed histologically to have LN lesions. Of the samples (n = 79) collected in parallel from apparently unaffected (HU) areas on the same diseased colonies showing gross signs of SCTLD across species, a prevalence of 11% were confirmed histologically with LN lesions ( Table 1). Of the apparently healthy (HH, n = 41) colonies evaluated from SCTLD sites, a prevalence of 15% had LN lesions. Samples collected from healthy (reference) sites (HH, n = 51) did not show LN lesions, except one M. cavernosa from Dustan Rocks (April 2017), in which early-stage LN was observed (Table 1; Figures 6A,B,D).

DISCUSSION
The histopathology findings presented here indicate that SCTLD is a bottom-up process starting with pathological changes in BBW and SBW zooxanthellae with lesions first appearing in the gastrodermis of the BBW. In multiple coral species, lesions progress to loss of the BBW tissue followed by attendant tissue loss in adjacent SBW, leading to sloughing and a gross lesion of acute, subacute, or chronic tissue loss. Feasibly, the advancement of surface lesions and their progression at different rates of tissue loss in various coral species could be caused by secondary pathogens or opportunists that colonize following initiation of BBW LN lesions by still undetermined primary causal agents. Bacteria apparently play a role in SCTLD progression, as evidenced by the slowing of lesion progression after the application of antibiotics . While Vibrio coralliilyticus toxins have been detected in ∼20% of M. cavernosa with SCTLD 7 , we saw no evidence that bacteria are the primary cause of SCTLD.
The consistency of the LN lesion across the scleractinian species examined here seems to be a unique phenomenon in coral TLDs, which, at least in the Pacific, present with myriad microscopic manifestations (Williams et al., 2011;Work and Aeby, 2011;Work et al., 2012;Rodríguez-Villalobos et al., 2015). The presence of lesions in the BBW of even apparently healthy corals also indicates that LN starts before the manifestation of gross lesions. This could have important ramifications for management protocols that assess the health of corals based on their gross visual appearance alone and may require additional action (e.g., quarantine) before corals can be approved for transplantation, rescue, or coral-restoration purposes (Hein et al., 2017). Specifically, even apparently normal-appearing corals may not be healthy, and additional tests including, at least, microscopic examinations of tissue might be necessary in evaluating whether corals are SCTLD-free prior to such management actions. The possible early LN lesion from an M. cavernosa sample collected in April 2017 from a reference site in Dustan Rocks could be a significant finding, as it would place SCTLD in the Middle Keys months earlier than has been documented. If this sample does indeed represent SCTLD (although further screening is required of "healthy" colonies in SCTLD-endemic zones), then the development of diagnostic tests for SCTLD in corals in which no surface lesions are evident is further merited.
Tissue loss diseases, including SCTLD, in the Caribbean are increasingly notorious for causing coral declines Neely and Lewis, 2020). During the early appearance of SCTLD in the Florida Reef Tract (Precht et al., 2016;Gintert et al., 2019), researchers considered that this disease was the same as or similar to WP (or variants) reported in mostly the same coral species in Florida (Dustan, 1977;Richardson, 1998;Richardson et al., 1998a,b;Porter et al., 2001) and in the Caribbean, where ∼ 40 species have been affected (Green and Bruckner, 2000;Weil, 2004). Another grossly similar disease to SCTLD, reported in June 2008 from the Dry Tortugas (Tortugas multispecies rapid TLD [TMRTLD], Brandt et al., 2012), affected similar scleractinian species. However, absent systematic histopathology descriptions, it is difficult to compare SCTLD to these other TLDs, illustrating the confusion that can arise when coral diseases are described based only on gross lesions (Work and Aeby, 2006).
Based on TEM observations in more than 80% of the WP tissue grids examined, Borger (2003) described LN below the epidermis in C. natans and S. siderea and suggested that the lesions might be caused by toxicoses associated with bacterial or fungal infections, associated anoxia, or endo-or exotoxins. Additionally, in a brief report showing histologic lesions of presumed WP in Dichocoenia stokesii from Navassa Island, in the Caribbean, and O. faveolata from Puerto Rico, Galloway et al. (2007) revealed diffuse gastrodermal necrosis that appeared to be markedly similar to the LN lesions described here. In a historical context, and to assess whether SCTLD has been documented (as WP or TMRTLD) on a smaller scale since the late 1970s, unraveling the relationship between these diseases is relevant. Assuming that SCTLD somehow originated in 2014 may be misleading if the disease is instead the re-emergence of a previously documented TLD.
Although SCTLD can manifest with varying rates of tissue loss, colonies of most species show a noticeable area of recent necrosis (indicated by intact bare, white skeleton) following the tissue-loss margin before overgrowth by turf algae on the exposed skeleton. Grossly, S. siderea colonies displayed a characteristic pink discoloration that can be challenging to discern from dark spot syndrome (Santavy et al., 2001) or other undescribed lesions. In Siderastrea spp. showing marked variability in pink pigmentation associated with bleaching, Sassi et al. (2015) reported that pocilloporin fluorescent pigments became visible in the deeper tissue layers with reduced zooxanthellae densities (as described by Bongiorni and Rinkevich, 2005). Perhaps similar mechanisms exist for S. siderea with SCTLD.
Given that SCTLD is a breakdown of host-symbiont relationships, understanding whether the process starts with the symbiont or the coral host cells seems of paramount importance. Feasibly, different genera of Symbiodiniaceae may be more susceptible to SCTLD than others and if so, this may explain the apparent range of coral species affected, with some species unaffected to others being highly affected. Functional Symbiodiniaceae species differences across tissue compartments have not been assessed and perhaps zooxanthellae located deeper in tissue, with less access to light (or sensitive to other abiotic or biotic factors), are more susceptible to whatever etiology is causing SCTLD. In summary, the fact that zooxanthellae are initially affected provides an important clue.
A plausible cause of SCTLD, based on light microscopic findings, is toxicosis. Perhaps toxins selective to zooxanthellae or host cells are responsible for symbiont breakdown. For instance, free-living dinoflagellates produce diverse toxins and bioactive compounds with an array of effects on aquatic animals (Landsberg, 2002). Toxin production can be intermittent and driven by abiotic factors (e.g., temperature, nutrients) or biotic factors (e.g., allelopathy, predation) (Cembella and John, 2006). Symbiotic dinoflagellate zooxanthellae also produce secondary metabolites (e.g., zooxanthellatoxins) that have diverse bioactivities in vitro (Nakamura et al., 1993;Fukatsu et al., 2007;Beedessee et al., 2015Beedessee et al., , 2019, but little is known of the function of these natural compounds in their coral hosts (Gordon and Leggat, 2010) or of any negative effects they may have (McConnaughey, 2012). Plausibly, zooxanthellae could become toxic to the host, leading to the initiation of SCTLD. Lytic necrosis is reminiscent of lesions caused by bioactive lytic compounds, such as proteases, more typically associated with bacterial pathogens (Miyoshi and Shinoda, 2000;Newman and Clements, 2008). The possible role of biogenic toxins or allelochemicals from other coral-associated organisms warrants consideration.
Zooxanthellae also produce various metabolites that can disrupt symbiosis. In response to environmental stress or exposure to pathogens, increased production by zooxanthellae of, for example, reactive oxygen species (ROS), nitric oxide, antioxidants such as dimethyl-sulfoniopropionate, and volatile organic compounds could play a role in the death of zooxanthellae and host coral cells (Mydlarz and Jacobs, 2004;Bouchard and Yamasaki, 2008;Mydlarz et al., 2009;Hawkins and Davy, 2012;McGinty et al., 2012;McLenon and DiTullio, 2012;Caruana and Malin, 2014;Wietheger et al., 2018;Lawson et al., 2019;Zhou et al., 2019). As part of their innate immune response against pathogens, corals also produce antioxidants such as melanin, with synthesis activated by cleavage of prophenoloxidase to phenoloxidase (PO) (Palmer et al., 2008). ROS produced during this process can be cytotoxic to pathogens but can also cause self-harm unless countered by antioxidants (Wright et al., 2017). In diseased and healthy colonies of Acropora millepora with the TLD white syndrome, increased PO activity was detected at the lesion border (Palmer et al., 2011). The putative detection of melanin by FM stain, shown here (Figures 7E,F) in degrading zooxanthellae closely associated with LN lesions, might indicate a pathological response to excess ROS, but more research would be beneficial.
Potential pathogens of zooxanthellae may also be causal factors. Bacteria (Ritchie, 2011;Bernasconi et al., 2019) and viruses in zooxanthellae have been documented (Wilson et al., 2005;Lohr et al., 2007) and associated with bleaching (Correa et al., 2016) and TLDs (Lawrence et al., 2014(Lawrence et al., , 2015Soffer et al., 2014), but Koch's postulates have yet to be met (Thrusfield, 2016). Initial evaluation by TEM of zooxanthellae from Montastraea cavernosa in this study revealed no discernible viral particles. But further ultrastructural examination of corals is clearly needed if viruses are to be ruled out as a cause of SCTLD.
Zooxanthella immune dysfunction may also be a significant contributory factor to SCTLD. For example, plants have a wellcharacterized immune response, but upon exposure to certain insults their hypersensitivity response leads to cell death (Jones and Dangl, 2006;Coll et al., 2011;Leary et al., 2018). Another mechanism, one whereby the host-symbiont relationship is disrupted, could be operating if zooxanthellae can suppress host immunity (Merselis et al., 2018) or if the host cell no longer recognizes the endosymbiont as self, leading to degradation and death of zooxanthellae. Sorting these processes out, however, would require much deeper understanding of coral host-cell immunity, a field still in its infancy (Mydlarz et al., 2006(Mydlarz et al., , 2016Palmer and Traylor-Knowles, 2012).
Finally, the Florida coast has had a long history of watershed alteration and pollution, with significant effects on corals (Rogers, 1990;Marubini and Davies, 1996;Lipp et al., 2002;Downs et al., 2005;Wagner et al., 2010). Associated factors, such as contaminants, pesticides, toxicants, nutrients, and sedimentation, may selectively kill or stress zooxanthellae (Owen et al., 2003;Reichelt-Brushett and McOrist, 2003;Brodie et al., 2012), but their role, if any, in coral disease has been less well studied. All these factors would be fruitful avenues of investigation.
Although the spatial epidemiology patterns  and the apparent efficacy of antimicrobial treatments on lesions  indicate that SCTLD is infectious, no evidence was visible at the light microscopic or ultrastructural level of infectious agents consistently associated with symbiont-host disruption. Although pathogenic microorganisms may not necessarily be observed histologically, the assumptions based on previous studies that Koch's postulates have been fulfilled by repeatedly isolating and culturing the same candidate microorganism from a lesion, recreating the lesion experimentally with that organism, and then re-isolating the microorganism (Rosenberg, 2004), are often done in the absence of histopathological confirmation of the primary pathogen in co-associated early lesions (Arboleda and Reichardt, 2010). Lesions can be recreated with pathogenic bacteria, but they do not necessarily prove they cause the same disease, especially if as in corals, many TLDs grossly appear the same (Work and Aeby, 2006). That some diseases have not subsequently been recreated, nor the purported primary pathogen again detected, has also led to confusion as to the correct etiology of some coral diseases (Sunagawa et al., 2009a;Sutherland et al., 2016).
Additionally, an obvious host-cell inflammatory response (e.g., amebocytes) was not seen in histopathology of SCTLD, as might have been expected if corals were targeting pathogens . Thus, while the identities of the coccoidlike and coccobacilloidlike structures need to be confirmed, bacteria, fungi, and parasites can initially be ruled out as proximate causes of SCTLD. But not all virus-induced lesions at the microscopic level will show viral particles (Russell et al., 2008), and a role for viruses in affecting either the corals or their zooxanthellae remains a possibility. Indeed, the commonality among the multiple coral species susceptible to SCTLD may lie not with the corals themselves but with their associated zooxanthellae, which are less diverse than affected corals at the genus level.
Although SCTLD can be recreated experimentally  this does not necessarily lead to the conclusion that the disease is caused by an infectious microorganism, although exposure of healthy coral fragments to an SCTLD-affected fragment elicits tissue loss, indicating that a transmissible pathogen is involved. Perhaps corals with SCTLD experimentally transmit "non-self " organisms or allelochemicals that induce a host response that initiates lesion formation. Secondary opportunistic bacteria can still play a role in lesion advancement. Finally, although this seems less likely because of documented experimental transmission, SCTLD may result from an environmental co-factor or cause, as postulated with sea star wasting disease (SSWD). Originally thought to be transmissible and infectious based on field and molecular data, no compelling link between viruses and cell death has been made at the cellular level, except possibly in one sea star species (Hewson et al., 2014(Hewson et al., , 2018, and at least 10 different densoviruses (primary suspect pathogen) have been found to be equally represented in symptomatic and asymptomatic sea stars (Jackson et al., 2020). Although the etiology is still unknown, one possibility is that SSWD is caused by temperature anomalies or other environmental factors (Hewson et al., 2018).
If the BBW gastrodermis is the first compartment to be affected in SCTLD, understanding how this process is initiated may be difficult to discern. In contrast to zooxanthellae, much less is known about the cell types that make up tissue compartments in corals. Although the basic layers (calicodermis, gastrodermis, mesoglea, epidermis) have been well documented, as are certain cell types such as cnidae (Peters, 2016), little is known about the complex physiological interactions between zooxanthellae and their host cells that would result in the pathology changes seen here. Nor is it clear whether initiation of LN in the BBW might be driven by influences from the calicodermis-skeleton interface or the GVC. For example, various macromolecules involved in skeletal calcification or digestion are enzymatic, lytic, or possibly toxic (Ramos-Silva et al., 2013;Deutekom et al., 2016;Raz-Bahat et al., 2017). Development of animal models for diseases of cnidarians might help in understanding fundamental cellular processes, for example, models utilizing Exaiptasia pallida Sunagawa et al., 2009b) or Hydra viridissima (Kovacevic, 2012).
Other changes in corals in the present study were considered secondary. The CIBs found only in M. cavernosa and P. strigosa were inconsistently associated with LN lesions in the BBW, and their function is unknown. Diverse CIBs (e.g., organic crystalline metabolites) are often found in normal organisms (Edwardson and Christie, 1978;Nürnberger et al., 2017) and while the CIBs reported here are similar in ultrastructural appearance to the kalisomes in the gastrodermis of coral larvae (e.g., Pocillopora damicornis) (Clode and Marshall, 2002), they could be crystalline inclusions resembling aragonite calcium carbonate (Hayes and Goreau, 1977), and still have to be analyzed chemically. When crystallized in the laboratory at pH 8.0, the coral chromoprotein pocilloporin forms a mauve tetrahedron (Beddoe et al., 2003) reminiscent of the CIBs noted here, but it is not known whether this process would occur in corals in situ. Among other biological roles, pocilloporins are thought to be ROS scavengers (Bou-Abdallah et al., 2006;Furla et al., 2011), which could relate to the symbiont-host physiological dysfunctions mentioned previously. Finally, some of the parasites (e.g., Halofolliculina sp. ciliates) noted here in corals with SCTLD are likely incidental following skeletal exposure. Ciliates have been associated with TLD lesions in the Pacific (Work et al., 2012) and in the Caribbean (Verde et al., 2016). Of particular interest are the trichodinid ciliates not reported previously in corals and the epidermal apicomplexans. An apicomplexan coccidian in corals has been described (Upton and Peters, 1986), and, while other apicomplexans have been reported (Kirk et al., 2013;Kwong et al., 2019), whether they are pathogenic is unknown.
In summary, SCTLD may have a non-infectious primary etiology (with a likely role for secondary opportunists or pathogens), but much remains to be done to unravel the actual cause. Improving understanding of how the disease progresses among species by using histology to monitor changes over time might refine the case response for individual species, some of which appear more susceptible than others. Developing an animal model for coral diseases and immune function would also elucidate mechanisms of disease and responses to insults by both coral and symbiont. Additional ultrastructural and molecular studies might allow for the detection of viruses and better understanding of interactions between hosts and symbionts at different stages of disease. While scientists continue to understand the causes of SCTLD, broader management activities to reduce stresses to coral reefs such as reductions in land based pollution (Fabricius, 2005) could benefit coral reefs in Florida.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
JL secured funding. JL, YK, and EP designed the research. YK prepared samples for histology. PW, NP, and YW conducted histology and special staining. JL and YK performed TEM. KM and LH managed collections and field data. JL, YK, EP, and TW evaluated histological slides and interpreted histopathology. JL and TW drafted the manuscript. All authors contributed to writing the manuscript.

FUNDING
This research was supported by U.S. Fish and Wildlife Service Funds for State Wildlife Grant Award FL-T-F18AF00492 awarded to FWC and supplemented by funding to FWC for field collections from the Florida Coastal Office, Florida Department of Environmental Protection Coral Reef Conservation Program (order numbers BOA35A and B25636).