Manipulative aquaria-based experiments were utilized to complete the two disease-induction studies, referred to as the (i) M. cavernosa to M. cavernosa and (ii) M. cavernosa to O. faveolata experiments. The first induction experiment was conducted from April 26–June 25, 2019 using AH and diseased (subacute tissue loss) M. cavernosa. The AH M. cavernosa colonies were collected on April 23, 2019. The second induction experiment was conducted from July 26–September 26, 2019 using AH O. faveolata and diseased (subacute tissue loss) M. cavernosa. The AH O. faveolata colonies were collected on July 23, 2019. The AH colonies for each experiment were collected at the Airport Coral Heads site off Key West (24.53919°, −81.77270°), a site that was outside of the known distribution of SCTLD within FCR at the time of collection. Seven AH colonies, approximately 30 cm in diameter, of each species were collected for each experiment via hammer and chisel. Field participants wore sterile gloves during the collection process. After the colonies were removed from the reef, they were transported by hand up to the boat and stored in a sterile cooler with ambient seawater-soaked bubble wrap. The AH colonies were then transferred to an isolated aquarium containing ambient saltwater from a deep-water well at Mote Marine Laboratory’s (MML) International Center for Coral Reef Research and Restoration in Summerland Key, FL and maintained under those conditions for 1–2 days while the diseased corals were collected. AH colonies were collected prior to diseased colony collections to minimize contamination and inadvertent spreading of the disease. Original plans to expose AH O. faveolata to diseased O. faveolata were not possible because diseased O. faveolata colonies were not available at the time of collection. Therefore, diseased M. cavernosa colonies for the first and second experiments were collected on April 25, 2019 and July 24, 2019, respectively, in the Looe Key reef area (24.54767°, −81.45697°). These colonies exhibited grossly visible signs consistent with SCTLD, were approximately 30 cm in diameter, and had approximately 30% disease progression. Each colony had a diffuse subacute tissue loss lesion, defined as a lesion with bare white skeleton less than five centimeters in width along the lesion edge (described in Aeby et al., 2019). At the time of collection (for both experiments), this lesion type was the most prevalent in the area of collection. This lesion type was also used in previous SCTLD induction studies (Aeby et al., 2019). A total of eight diseased M. cavernosa colonies were collected (four for each experiment). After diseased colony collections were completed for each study, the AH and diseased colonies were transported from Summerland Key, FL to MML’s Coral Health and Disease laboratory in Sarasota, Florida for testing. The diseased and AH colonies were transported in separate coolers, wrapped with seawater-soaked bubble wrap. On arrival the diseased and AH colonies were placed in separate (previously bleached) bins with ambient seawater (∼102 L volume) obtained from the Coral Health and Disease wetlab supply source overnight. Each bin was filled with seawater, equipped with multiple powerheads (songlong SL-381 submersible pump) to maintain circulation, and temperature was regulated by a recirculating temperature-controlled water bath.

One 3 m × 1.2 m fiberglass raceway holding twenty 18.9 L aquaria was used for each experiment, to keep the temperature of the water within the aquaria constant. Temperature within the raceway was controlled by a heat exchanger that was set to a temperature of 27.5°C. Corals were held under a 10-h light: 14-h dark photoperiod using Radion XR30w Pro aquaria lights (EcoTech Marine, Allentown, PA, United States). The lights were set to a “traveling sunrise and sunset” feature, with fluctuating values of 50–300 PAR over the course of a day to mimic natural conditions as much as possible.

Using sterilized gloves, each diseased M. cavernosa colony (n = 4 × 2 experiments, identified as diseased donor colonies (DC) A to D [DCA–DCD]) was taken from the aquaria (∼24 h, including transportation time), and fragmented on a sterile table into four donor pieces, with each piece representing a sequential “distance” away from the active tissue-loss margin (Figure 1A). Coral colonies were fragmented using an angle grinder (Dewalt DWE402) and trimmed with a diamond blade band saw (Gryphon Corporation, C-40 CR Aquasaw XL). To minimize contamination, all blades and tools were sanitized with a 10% bleach solution between uses. Each diseased donor fragment in the induction experiments was approximately 16 cm long × 8 cm wide × 3–5 cm high. The diseased colonies were fragmented parallel to the lesion edge to determine whether isolated AH donor fragments that showed no obvious grossly visible tissue-loss signs, would subsequently exhibit tissue loss and induce SCTLD in recipient M. cavernosa or O. faveolata fragments. The donor fragment with active tissue loss was identified as “distance 1” (D1). The sequential donor fragments with no obvious grossly visible tissue-loss signs were approximately 4 cm (“distance 2” – D2), 12 cm (“distance 3” – D3) and 20 cm (“distance 4” – D4), respectively, away from the active tissue-loss margin (Figure 1A). One diseased M. cavernosa colony collected for the O. faveolata transmission experiment had multiple subacute lesions, therefore this colony (DCD) was fragmented into three D1 pieces only (i.e., no D2–D4 donors were used).

For the M. cavernosa to M. cavernosa experiment, one AH M. cavernosa colony was randomly chosen to act as a donor control and was cut into four fragments of similar size as the diseased donor colony fragments. All AH colonies were fragmented first before fragmenting each diseased donor colony (and each treatment fragment within the donor colony) to reduce the possibility of cross contamination. The remaining six AH M. cavernosa colonies were fragmented into approximately ten pieces each, ∼10 cm long × 3 cm wide × 2 cm high, to act as recipient fragments.

For the M. cavernosa to O. faveolata experiment, an AH O. faveolata colony was randomly selected as a donor control and was fragmented into four fragments of similar size to the diseased donor colony fragments. Additionally, four AH M. cavernosa fragments of similar size were placed into control aquaria to account for inter-species aggression. The remaining six AH O. faveolata colonies were fragmented similarly to the M. cavernosa to M. cavernosa study (Figure 1B).

For each experiment, the recipient fragments were randomly distributed among the aquaria with equal distribution to each treatment and the controls. Each aquarium had one diseased donor fragment of a different treatment with three recipient fragments directly touching it. Some of the D1 donor fragments had to be cut in half so that each recipient fragment could be in contact with the active tissue-loss margin (Figure 2A). There was an excess of recipient fragments, so these fragments were randomly placed in aquaria, but not in direct contact with the donor fragment to assess the potential for waterborne transmission (Figure 2B). These fragments were designated as ‘non-touching.’ Four of the 20 aquaria were set up as controls to account for any aggression from direct contact. Recipient fragments were placed in direct contact with the AH donor colony that was randomly chosen as a control. The experimental design matrices for the aquaria set ups for the two induction experiments are summarized in Figures 3A,B (see Supplementary Tables 1, 2 for details).

All seawater at MML was pulled from the local Sarasota Bay, and went through several filtration and ozonation steps before entering the laboratory. Fifty percent water changes were completed daily on all aquaria, and a recirculating water pump (songlong SL-381 submersible pump) was used to maintain water motion within each aquarium. All contaminated seawater along with any exposed equipment was bleached (1:9 bleach to water ratio) and UV sterilized prior to disposal. Water quality parameters (including temperature, pH, salinity, and dissolved oxygen; Supplementary Table 3) of the seawater utilized for the 50% water changes were monitored daily using a YSI handheld instrument (YSI ProDSS, Yellow Springs, OH, United States).

Aquaria observations were recorded and photographs were taken daily to monitor tissue loss appearance and progression within the donor fragments and the touching and non-touching recipient fragments. When recipient fragments exhibited more than 5% tissue loss, they were closely monitored and prepared for histology sampling.

Baseline samples were collected for histopathological examination from the diseased and control donor fragments for both experiments before donors were assigned to experimental aquaria. In addition, recipient M. cavernosa and O. faveolata fragments were sampled at time zero and through time as tissue loss was grossly observed. Samples were collected when fragments had approximately 25, 50, and 75% tissue loss (so the same fragment could have been sampled up to three times, unless the fragment reached complete tissue loss before the sampling time point or the tissue loss had stopped progressing). Each sample collected from a recipient fragment within a treated disease aquarium was paired with a sample collected from a control recipient fragment. Control samples were taken from fragments of the same parent colony (A–B, D–G) as the recipient fragment being sampled from the treated aquarium.

When a recipient fragment was ready to be sampled, the fragment was removed with sterile gloves from its aquarium, photographed, and a 3 cm × 3 cm × 2 cm sample was extracted using a sanitized diamond-blade band saw (Gryphon Corporation, C-40 CR Aquasaw XL). After sampling, the coral fragment was re-photographed and placed back into the same aquarium in its original location touching the donor fragment, unless it was a non-touching fragment.

All histological samples were fixed with a solution of 1 part Z-Fix (zinc formalin; Z-Fix concentrate [18.5% formaldehyde; Anatech Ltd., Battle Creek, MI, United States]) mixed with 4 parts 0.2 μm-filtered, UV-sterilized natural seawater. Samples collected are summarized in Tables 1–4 (see Supplementary Tables 4–7 for details). All collected samples for histologic examination were processed, except that a subset of the paired sampled controls were randomly selected. All processing for histology was conducted at the Florida Fish and Wildlife Conservation Commission’s Fish and Wildlife Research Institute, St. Petersburg, FL, United States.

Before histopathology analyses, all samples were photographed using a digital camera fitted with a macro lens (Nikon, Tokyo, Japan). Diseased fragment samples exhibiting grossly visible lesions were further examined under higher magnification with a dissecting microscope attached to an Olympus DP72 digital camera (Tokyo, Japan), and photomicrographs taken. At this point folliculinid ciliate (heterotrich) infections (Verde et al., 2016), especially Halofolliculina sp., in the skeleton were also determined. These organisms were not detectable by histology because all skeleton, including these organisms, is removed during decalcification.

Coral tissue for routine histological processing followed Landsberg et al. (2020). Briefly, tissue was enrobed with agarose (for samples exhibiting gross lesions only), decalcified with 10% ethylenediaminetetraacetic acid (EDTA, Fisher Scientific, Na₂⋅2H₂O; MW = 372.1), and then processed for routine histology. After all the aragonite skeleton had dissolved away, decalcified soft tissues were also photographed microscopically prior to processing. Since the fragments were large enough to make both sagittal and radial cuts, tissues were trimmed this way – sagittally (perpendicular to polyp mouth) and radially (parallel to polyp mouth), embedded in paraffin (Paraplast Plus, Fisher Scientific, Waltham, PA, United States) or glycol methacrylate plastic resin (JB-4; Electron Microscopy Sciences, Hatfield, PA, United States). Embedded tissue blocks were sectioned at 4.0 μm with a rotary microtome and stained with Mayer’s hematoxylin and eosin (H&E for paraffin medium), Weigert’s H&E (for JB4 medium), thionin, or periodic acid Schiff/metanil yellow (PAS-MY for JB-4; Quintero-Hunter et al., 1991). Slides were examined with an Olympus BX51 light microscope equipped with an Olympus DP71 digital camera (Olympus Inc., Tokyo).

An initial screening of three slides (two paraffin and one JB-4 embedded medium as described above) from each specimen was conducted to assess if slides were suitable for further examination. For example, if the histological slide was not well prepared or missed the target tissue area for lesion evaluation (gastrodermis of basal body wall, BBW), then the specimen was censored from the histological examination.

All statistical analyses were performed in R-4.0.2 (R Core Team, 2020), and figures were made using Prism (version 8.0.0 for Windows, GraphPad Software, San Diego, CA, United States). A Fisher’s Exact Test (Fisher, 1934) was used to examine the likelihood of tissue loss for the different treatment “distances” within the diseased donor colonies. The likelihood of the different treatment “distances” to exhibit tissue loss was examined for each experiment separately; and across both experiments.

A Fisher’s Exact Test was also used to identify statistical differences in the probability of touching vs. non-touching M. cavernosa recipient fragments to exhibit tissue loss, and a Chi-Square analysis (Pearson, 1900) was used to identify statistical significance of touching vs. non-touching O. faveolata recipient fragments to exhibit tissue loss.

A Fisher’s Exact Test was used to examine the likelihood of tissue loss in both M. cavernosa and O. faveolata recipient fragments exposed to different treatments. A Chi-Square analysis was used to compare the likelihood of tissue loss between species. A Fisher’s Exact Test was used to identify statistical significance of M. cavernosa vs. O. faveolata non-touching recipient fragments to exhibit tissue loss. Pearson’s correlation (Benesty et al., 2009) was used to assess the correlation between distance away from the lesion and time of tissue loss onset within the D2–D4 donor fragments.

Disease progression (i.e., tissue loss) within recipient-touching fragments was measured using ImageJ software (Abràmoff et al., 2004) on the photographs of the fragments taken in the aquaria during the experiment. Disease progression was measured only in recipient fragments that were affected by tissue loss that led to complete mortality of the tissue (i.e., all tissue had sloughed off of the skeleton). Disease progression was assessed by measuring the area of living tissue at the halfway point between initial grossly visible tissue loss signs and death (on average 3.5 days after initial tissue loss appearance for each species) and subtracting from the initial area of live AH (pigmented) tissue. One-way analysis of variance (ANOVA) tests were used to assess the statistical significance between species for timing of tissue-loss onset (i.e., disease induction) in recipient fragments among the treatments and lesion progression (cm²/day). Timing of tissue loss onset (i.e., disease induction) in recipient fragments among the treatments was measured after a grossly visible tissue loss lesion appeared on the donor colony. For the M. cavernosa experiment, this data was transformed using an aligned rank transformation (Conover and Iman, 1981).

A Cox proportional hazard model (Cox, 1972) was used to compare the survival times of M. cavernosa and O. faveolata touching recipient fragments. It was also used to compare survival times of touching recipient fragments among the four treatments within M. cavernosa and O. faveolata.

Across both experiments, six out of the seven (85.7%) D2 M. cavernosa donor fragments exhibited gross signs of SCTLD, and five out of the seven (71.4%) D3 and D4 fragments exhibited signs of SCTLD (Supplementary Figure 6).

When reviewing data for both experiments combined, the correlation between the distance from the lesion border and time of tissue loss onset was not significant (Pearson’s correlation, coefficient = 0.82, p = 0.18). After collection from the field, it took an average of 12.5 (±2.64) days for initial signs of tissue loss to appear within D2 fragments, 9.8 (±3.61) days for D3 fragments, and 17.6 (±2.94) days for D4 fragments (Supplementary Figure 7).

Results of this study indicate that tissue loss appears and progresses within isolated fragments even after their removal from the grossly visible lesioned area in diseased (subacute tissue loss) M. cavernosa colonies, approximately 30 cm in diameter. The likelihood of a donor fragment exhibiting tissue loss was not significantly different among the different donor disease treatment distances away from the active lesion, suggesting that SCTLD could be systemic within the colony. AH tissue in fragments taken as far as 20 cm from the grossly observable lesioned border area still became diseased even after fragment removal and isolation from the grossly visible diseased portion of the colony. It has been documented in the field that topical antibiotic applications are effective in halting SCTLD lesion progression (Aeby et al., 2019; Neely et al., 2020; Shilling et al., 2021); however, the effectiveness is likely to remain localized in the region of application (Neely et al., 2020), while new lesions can appear in other areas of the colony following treatment. Colony size may be a factor in lesion appearance among the different treatment distances. It is possible lesion appearance would not be observed grossly in the AH tissue directly adjacent to a subacute lesion within a larger colony, as new multifocal lesions typically appear throughout the colony with possible mortality occurring over several years in larger colonies of intermediately susceptible species (see Text Footnote 1). Additionally, the tissue could have already been diseased (sub-surface systemic lesions) upon collection but did not exhibit the grossly obvious appearance of SCTLD (active tissue loss and/or bleached tissue adjacent to the lesion border). Indeed, three donor colony samples showed LN within these sections of the colony at the time of collection and prior to experimentation. Results of Neely et al. (2020) and the present study suggest a colony-wide treatment may be necessary to effectively preserve remaining coral tissue. The present study suggest that SCTLD may be systemic within a 30 cm maximum diameter colony, but additional handling stress from transportation, fragmentation, or repeated sampling may have increased the susceptibility of the D2–D4 fragments to SCTLD in an experimental setting. Additionally, it is possible that tissue loss occurred in the D2–D4 fragments because there were other corals (i.e., recipient fragments) directly touching them. The D2–D4 fragments were already compromised before showing active tissue loss and the direct contact with recipient fragments could have exacerbated their condition. However, it is important to note that no control donor fragments experienced tissue loss in either experiment even though they were handled and treated similarly. Future studies should examine D2–D4 fragments in aquaria by themselves to assess whether tissue loss still occurs.

In the present study, the successful SCTLD induction under controlled laboratory conditions was confirmed by histopathologic examinations for the presence of LN in the BBW (Landsberg et al., 2020) for recipient fragments in direct contact with donor fragments. LN was not confirmed in recipient fragments that did not have physical contact with a diseased fragment, although only a few of these samples were examined histologically. None of the recipient fragments exhibited tissue-loss lesions when donor treatments (D2–D4; i.e., at least 8 cm away from the grossly obvious SCTLD lesion) did not exhibit grossly visible SCTLD lesions. This suggests that it is necessary for the diseased donor fragments to exhibit grossly visible SCTLD lesions to induce disease in recipient fragments by physical contact. However, we did not evaluate by histological examination any recipient fragments that did not exhibit SCTLD lesions at the grossly visible level. Control recipient fragments were examined in parallel and showed LN histologically on two fragments (Table 2) but did not progress to SCTLD grossly visible lesions. Also, there were several recipient fragments that experienced disease induction but the active tissue loss stopped during the study. This may have been from a disruption to the disease etiology from sampling, as all of these recipient fragments were sampled at least once before the lesion stopped progressing.

The histological applications were a useful tool to couple with grossly visible tissue loss within this study, but it would have been useful to confirm the lack of LN in the BBW prior to the study beginning to ensure recipient corals were indeed healthy visibly and microscopically. However, histological processing involves time-consuming laborious work that includes decalcifying the skeleton. Histopathologic examination is a vital tool, but it may be important to develop more rapid, robust diagnostic tools for evaluating coral health status.

For both experiments, induction of tissue loss occurred on recipient fragments touching D2–D4 donor corals. However, the D2–D4 donor corals had to develop a grossly visible tissue-loss lesion in order for induction to occur. This has been documented in other transmission studies (Aeby et al., 2019; Meiling et al., 2021) — donor corals within these studies had a grossly visible tissue loss lesion when induction occurred in both touching and non-touching recipient fragments. However, this is the first study to document that the apparently healthy tissue adjacent to a grossly visible lesion can develop a lesion and ultimately induce disease in apparently healthy corals. Additionally, there was no significant difference in the average number of days prior to tissue-loss signs in recipient-touching fragments among those exposed to the different distances. Although tissue loss within recipient-touching fragments was observed less in further-distance exposures, on average it took the same amount of time for initial signs of tissue loss to appear within the recipient-touching fragments (after a grossly visible tissue-loss lesion appeared on the donor colony).

For both experiments, there was no significant difference in the gross appearance of tissue loss for touching and non-touching fragments. However, not all fragments were examined histologically for LN. Successful SCTLD induction was confirmed with histopathologic examinations by the presence of LN at the BBW of the gastrodermis only for touching recipient fragments. One non-touching O. faveolata recipient fragment was putatively confirmed with LN, but data were equivocal as the tissue on the histology section was minimal with only one small putative BBW LN lesion observed. A previous study by Aeby et al. (2019) indicated that 30% and 10% of M. cavernosa touching and non-touching fragments, respectively, exhibited tissue loss when placed in an aquarium with M. cavernosa colonies showing sub-acute tissue loss. In contrast, 100 and 60% of touching and non-touching M. cavernosa fragments, respectively, experienced tissue loss when exposed to Colpophyllia natans with acute tissue loss lesions. The present study showed that 39 and 10% of touching and non-touching M. cavernosa fragments and 56 and 35% of touching and non-touching O. faveolata fragments, respectively, developed tissue loss from diseased M. cavernosa. This supports that touching and non-touching fragments show similar rates of induction when exposed to diseased (subacute tissue loss) M. cavernosa.

Biotic factors possibly affecting the controlled laboratory experiments also need to be considered. It is interesting to note that there were only two cases of ciliate infections in the M. cavernosa to M. cavernosa experiment, but 26 cases (including 4 controls) were found in the M. cavernosa to O. faveolata study. Halofolliculina sp. was first detected at 32 days post experimental start for both experimental trials. This suggests that the Halofolliculina sp. propagated during the test period after 32 days and probably O. faveolata is more susceptible or weakened by experimental conditions compared to M. cavernosa. It may be possible that the M. cavernosa to O. faveolata study was affected by ciliate infections. Additionally, tissue loss occurred similarly for all distances in O. faveolata, and there was no significant difference in the amount of time for first signs of tissue loss to appear in touching fragments among the four distances. These results suggest that the visual signs and etiology (i.e., disease ecology) of the new lesions that appeared on the distances away from the disease lesion were comparable to the active lesion on coral collections. However, artificial laboratory conditions might have encouraged the proliferation of a ciliate coinfection that would have altered the probability of disease occurrence and rates of progression.

Our results indicate that tissue loss occurred more often in O. faveolata compared to M. cavernosa, suggesting O. faveolata was more susceptible to SCTLD. Similar results have been documented in other SCTLD transmission studies (Aeby et al., 2019; Meiling et al., 2021); however, we did not detect any statistically significant differences between species in the present study. It is possible that concurrent infections by other organisms as discussed above may influence disease susceptibility among species. Also, different SCTLD pathogens or strains could affect M. cavernosa, which, in turn, could influence the susceptibility of O. faveolata to disease induction. We do acknowledge that the lack of diseased O. faveolata was a limitation for this study. However, this was unavoidable because there were no such colonies available in the field at the time of collection. Future studies should include an O. faveolata-O. faveolata trial to account for potentially different disease etiology between these two coral species. Additionally, it is important to note that these induction experiments were completed in an indoor controlled laboratory setting with artificial light conditions, and corals were fragmented immediately prior to exposure. Aeby et al. (2019) and Meiling et al. (2021) took place in outdoor systems, and Meiling et al. (2021) allowed for a 1-week acclimation period after fragging. Both of these studies showed Orbicella spp. had a greater susceptibility to SCTLD than M. cavernosa, similar to the present study. Regardless, results may differ under the influence of environmental factors that corals would encounter under natural conditions.

On average there was the same amount of time between initial disease appearance and complete mortality of both M. cavernosa and O. faveolata recipient-touching fragments. However, disease progressed significantly faster in M. cavernosa recipient-touching fragments compared with O. faveolata recipient-touching fragments. Interestingly, O. faveolata was more susceptible to SCTLD induction, but once disease was initiated, M. cavernosa lost tissue at a faster rate. These results have been observed in other studies. For example, Aeby et al. (2019) documented that M. cavernosa fragments touching diseased M. cavernosa developed lesions over a 2–6-days period, whereas O. faveolata developed lesions over a 2–12 days period. Meiling et al. (2020) measured absolute areal tissue loss rates in M. cavernosa and Orbicella annularis; and documented that M. cavernosa had a significantly higher rate of tissue loss when compared to O. annularis. This study observed average tissue loss of ∼7 cm²/day in M. cavernosa, and ∼2 cm²/day in O. faveolata, which is comparative to the present study (∼3 cm²/day in M. cavernosa, and ∼1.5 cm²/day in O. faveolata). Additionally, it is important to compare the disease induction susceptibility and lesion progression with species that are not considered to have intermediate susceptibility (i.e., high susceptibility and low susceptibility). Aeby et al. (2019) documented no disease induction when exposing a species with low susceptibility, Porites astreoides, to diseased (subacute tissue loss) M. cavernosa. They also documented 100% disease induction when exposing a species with high susceptibility, Meandrina meandrites, to a diseased coral; although the diseased coral was a different species with an acute tissue loss lesion.