Changes in global climate have expanded the range of many infectious diseases while the added stress from increased natural disasters, habitat loss, thermal extremes, and anthropomorphic impacts have left many species more vulnerable to infection by pathogens old and new (Daszak et al., 2001; Omazic et al., 2019; Price et al., 2019). As a result, many organisms have been experiencing an increased frequency and spread of disease outbreaks worldwide (Bartlow et al., 2019). Beginning in the summer of 2013, an outbreak of Sea Star Wasting (SSW) began decimating sea star populations along the west coast of North America. SSW disease has come to be recognized as one of the largest marine epizootics ever observed due to its wide geographic range from Mexico to Southern Alaska, and its impact on over 20 different asteroid species (Hewson et al., 2014). Cases of SSW can be traced back as early as 1896 in the eastern US (Mead et al., 1898), and small-scale wasting events have been recorded on the west coasts of North America since the 1970’s (Menge, 1979; Dungan et al., 1982; Jangoux, 1986; Eckert et al., 2000; Bates et al., 2009). However, none of these past events compare to the scale of wasting seen since 2013. SSW is of critical ecological concern as it threatens some of the most important keystone predators, resulting in large-scale impacts on biodiversity, community structure, and loss of essential habitats, such as kelp forests, which are home for many organisms and serve as nurseries for important fisheries (Menge et al., 2016). The sunflower star, Pycnopodia helianthoides, is among the most heavily impacted asteroids and has virtually disappeared from its native ranges from California to Oregon (Harvell et al., 2019). Now critically endangered, the remaining populations of P. helianthoides take refuge in northern Canada and Alaska (Harvell et al., 2019).

The etiology of SSW remains unknown. A wide array of signs commonly observed in wasting sea stars include loss of body turgor, ectodermal discoloration, puffiness, limb autonomy (twisting and curling), body wall lesions, and disintegrating tissue (Hewson et al., 2014). The disease progresses and spreads quickly, with death occurring in a matter of days and outbreaks can spread to encompass thousands of miles within months (Harvell et al., 2019). Some studies have suggested transmission through a water-borne infectious agent spreading from infected to healthy animals sharing the same aquaria (Hewson et al., 2014). In many cases, warmer ocean temperatures have coincided with wasting events (Bates et al., 2009; Staehli et al., 2009; Harvell et al., 2019), and cooler temperatures have been shown to improve asteroid survival (Kohl et al., 2016). However, cooling water temperatures coincided with increased wasting in one study off the coast of Oregon, suggesting that temperature alone is not the sole driver of SSW outbreaks (Menge et al., 2016). A controversial suggestion posits that SSW is not caused by a single causative agent, but a combination of factors including environmental change, abiotic stress (Aalto et al., 2020), undefined pathogens, and dysbiosis of the organisms’ microbiome (Lloyd and Pespeni, 2018; Aquino et al., 2020). Although early investigations implicated a densovirus as a plausible agent of the disease (Hewson et al., 2014), later studies failed to replicate these findings in other affected asteroid species in addition to the presence of the densovirus in apparently naïve individuals (Jackson et al., 2020).

Recent studies have begun investigating changes in microbial communities in association with SSW disease (Lloyd and Pespeni, 2018; Aquino et al., 2020). Host-microbe interactions play a crucial role in the development, nutrient acquisition, behavior, and pathogen resistance of most multicellular organisms (Peixoto et al., 2021). The immune system and the microbiome function synergistically to maintain homeostasis and deter pathogens; compromising one often results in the collapse of the other and may lead to immune dysregulation and increased susceptibility to infection (Peixoto et al., 2021). This disruptive process, termed microbial dysbiosis, can be broadly defined as any changes in the composition and/or function of resident microbial communities relative to those found in healthy individuals. This may result from pathobiont expansion, loss of microbial diversity, and/or loss of beneficial microbes (Petersen and Round, 2014).

Infections of microbial pathogens have been previously linked to a wide range of mass mortality events afflicting echinoderms (Scheibling, 1986; Tajima et al., 1996; Scheibling and Hennigar, 1997; Tajima et al., 2007; Becker et al., 2008; Delroisse et al., 2020). Alterations in microbial community composition have been recorded with disease onset and progression in the ochre sea star, Pisaster ochraceous (Lloyd and Pespeni, 2018). The proliferation of anaerobic bacteria at the animal-water interface, as a result of increased primary production and organic matter buildup during following warmer temperatures, has also been implicated as an early mechanism of disease onset by limiting dissolved oxygen supply to infected asteroids (Aquino et al., 2020). While the impact of SSW on the microbiome has been investigated in controlled laboratory conditions (Lloyd and Pespeni, 2018; Aquino et al., 2020), the function of the microbiome in natural marine systems remains understudied. Monitoring of natural communities and their associated microbiomes is critical to understanding the resilience of these communities in the presence and persistence of the recent wasting outbreaks. Investigations into these interactions, in both healthy and diseased organisms, will set the foundations for long-term investigations into the evolutionary and adaptive potential of the microbiome to environmental change at a scale that is difficult to replicate in laboratory settings (Leray et al., 2021).

In this study, we analyze the microbiome of the sunflower star, Pycnopodia helianthoides, as a factor of apparent health status and exposure at multiple field sites. Here, we collected tissue biopsies from wild sea stars sampled at seven field sites in Southeast Alaska when the disease was just starting to emerge in 2016 ( Figure S1 ). At the time of collection, some impacted sites were completely overtaken by the disease, with all sea stars showing signs of wasting, while less severely impacted sites had a mixture of wasting and apparently healthy asteroids. Apparently healthy sea stars were collected from sites where wasting was observed (Exposed), alongside animals actively showing signs of wasting (Wasting). Sea stars were also collected from locations yet to show any signs of wasting (Naïve), however, there was no way to certify the absence of the disease-causing agent due to its unknown etiology. To better understand differences between a naïve, exposed, and actively wasting microbiome, we examined patterns of diversity within and between samples, identified differentially abundant taxa that may play a role in disease progression, and inferred metabolic pathways enriched or depleted between compared groups. Specific communities of samples and microbial taxa with similar patterns of abundance were then found using a Bayesian network clustering analysis based on hierarchical stochastic block models, a statistically rigorous alternative to traditional clustering approaches which were recently used to study host-microbiome interactions in the human gut (Cobo-López et al., 2022). Together these analyses provide several complementary perspectives in support of dysbiosis as a mechanism of SSW disease and its onset.

Microbial communities play an important role in maintaining the health of their hosts by providing protection against harmful pathogens and aiding in the metabolism of organic compounds (Peixoto et al., 2021). Changes to the core microbiome, defined as any set of commonly occurring microbial taxa, or functional attributes associated with those taxa, can leave the host susceptible to disease (Neu et al., 2021; Peixoto et al., 2021). The goal of this study was to compare the microbiomes of sea stars completely naïve to SSW to those recently exposed and actively afflicted by the disease in the field. Though our analysis, we were able to identify 1) a pre-symptomatic increase in microbial diversity, 2) an increase in abundance of facultative and obligate anaerobes (most notably Vibrio) accompanied by changes in metabolic function, and 3) a consistent co-colonization of taxa in large numbers among impacted individuals, as well as a shift in the less abundant, more stochastic communities found in Naïve samples to entirely different communities in Exposed and Wasting samples. Our results reveal changes to the core microbiome of P. helianthoides preceding physiological signs characteristic of SSW and thus supports the hypothesis that an early dysbiotic event plays a key role in disease progression of SSW in the field, particularly in the proliferation of anaerobic taxa.

The initial role of host microbiomes in SSW was characterized by Lloyd and Pespeni 2018, demonstrating the progressive shifts in microbial community composition through the stages of wasting. Further studies on the mechanisms of these shifts were proposed by Aquino et al. in 2021, hypothesizing that increases in organic matter (such as algal blooms, nutrient runoff, or pollutants) could alter the interactions of microbes at the asteroid-water interface, and lead to copiotrophic proliferation, appearance of facultative anaerobes, and eventual colonization of strictly anaerobic taxa which may compromise host respiratory and immune functions. Our findings in these field samples further support these current theories in the etiology of SSW by revealing an early dysbiotic event characterized by a consistent suite of facultative anaerobes (i.e.,Vibrionaceae, Moritellaceae, Colwelliaceae) (Imhoff, 2005; Bowman, 2014; Urakawa, 2014) across Exposed individuals followed by an increase of obligate anaerobes (i.e.,Clostridiales, Fusobacteriaceae) (Olsen, 2014; Stackebrandt, 2014) and other opportunistic taxa in the Wasting asteroids ( Figure 2 ; Figure 3 ). Like all previous studies on SSW, except one (Hewson et al., 2014) which was later revised (Hewson et al., 2018), we did not find evidence for a single causative agent responsible for outbreaks of SSW, however, the unknown Vibrio species showing a 1200-fold increase with exposure is worth further investigation. Our results and others suggest that the cause of SSW is likely a complex interaction between environmental stressors, undefined pathogens, and changes in the host-immune system that lead to a dysbiotic microbial community perpetuating SSW disease.

The Anna Karenina principle of animal microbiomes (Zaneveld et al., 2017) states the microbial community composition of diseased individuals varies more than healthy individuals. In the present study, comparing Naïve to Exposed to Wasting individuals, opportunistic microbiota increased with each stage in a progressively stochastic manner. For exposed and wasting sea stars, there was an increase in taxa diversity and dissimilarity between samples with SSW exposure and onset of signs. Exposed sea stars showed marked changes in community composition even before signs of the disease were present. However, what is most interesting is how these communities changed. The increase in microbial richness (alpha diversity) from Naïve to Exposed groups along with the increase in heterogeneity in microbial composition between individuals of the same group (within-group beta diversity), indicates an initial recruitment or proliferation of opportunistic taxa across Exposed individuals before symptom onset. The initial shifts in microbiota in Exposed asteroids may be a stepping-stone leading to the higher diversity observed between members of the Wasting group, indicating a more opportunistic colonization and a stochastic progression of dysbiosis coinciding with the onset of disease as anticipated by the Anna Karenina principle of dysbiotic states.

Some variation in microbiome composition may be expected from the geographic distance between Naïve and Impacted sampling sites, which were approximately 70km apart ( Figure S2 ). However, recent studies on the structure of coral microbiomes suggest that host species have the largest influence on microbial composition, demonstrating relative stability across space and time within a host (Dunphy et al., 2019). Additionally, we found greater similarity between Naïve and Exposed samples (both apparently healthy) than we did between Exposed and Wasting (both at impacted sites) within our PCoA plot ( Figure 1B ), supporting the idea that disease state, not geographic distance, is driving the observed variations in microbial community structure.

Further evidence of dysbiosis was found in the form of consistent co-colonization of taxa appearing in large numbers among impacted individuals ( Figure 3 , level-1, block 25; Table S2 ), forming a backbone community unique to the impacted samples. Unsurprisingly, the taxa in this block consisted of some of the most differentially abundant facultative anaerobes to appear between Naïve and Exposed samples in our microbial abundance comparison, including Vibrionaceae (Vibrio sp1., Photobacterium sp. and Aliivibrio sp.), genus Moritella sp., Shewanella sp., and Pseudoalteromonas porphyrae ( Figure 2 ). This suggests a shared set of microbes that may be important in the initial stages of SSW disease in exposed animals. Additionally, the community of microbes occurring both highly abundant and consistent across Naïve samples ( Figure 3 , level-1, block 26) seems to be less common across impacted samples. Members of this “healthy” backbone include taxa with the largest observed decreases in our differential abundance comparison including some species of Spirochaetaceae, Flammeovirgaceae, both unidentified bacteria phyla, and two members of the order Rickettsiales ( Figure 2 , Figure 3 , Table S2 ). This “healthy” backbone could be thought of as the core-microbiome of Naïve sea stars, and its reduced presence across Exposed and Wasting animals further highlights the microbial imbalance caused by SSW.

Of all the taxa that increased with disease exposure and symptom onset, members of the genus Vibrio are of particular interest due to their historical role in the pathology of many marine animals, including echinoderms, strong differential abundance trends ( Table 1 , Figure 2 ), and presence of multiple species in the backbone community of Impacted samples in our clustering analysis ( Figure 3 , block 25 Table S2 ). Previous studies have demonstrated the proliferation of Vibrionaceae species throughout the progression of SSW (Lloyd and Pespeni, 2018; Aquino et al., 2020), have been regularly identified in the decaying tissue of wasting sea stars prior to the current 2013 outbreaks, and associated with many other echinoderms diseases (Eckert et al., 2000; Becker et al., 2004; Staehli et al., 2009; Kohl et al., 2016; Hira and Stensvåg, 2022). Several species of Vibrio are known to cause intestinal and extraintestinal infections in humans and animals (Farmer and Hickman-Brenner, 2006), V. coralliilyticus is known to cause tissue damage and necrosis in corals (de O Santos et al., 2011), and V. echinoideorum was recently found to facilitate the development of lesion syndrome in the green sea urchin (Hira and Stensvåg, 2022). Interestingly, NCBI Blast results placed V. echinoideorum with the highest total-score for of the top 5 ASVs with the largest fold-change when comparing Naïve to Exposed sea stars. Additionally, a 2011 study focusing on population control methods of the Crown of thorns Sea Star, Acanthaster planci, used a thiosulfate-citrate-bile-sucrose agar (TCBS) to selectively stimulate the growth of Vibrionaceae species (Rivera-Posada et al., 2011a; Rivera-Posada et al., 2011b). Not only did the TCBS stimulation of Vibrionaceae species elicit SSW-like signs and mortality, but a follow up study in 2012 demonstrated an interspecies transmissibility of the TCBS-induced disease between different species of asteroids (Caballes et al., 2012). Taken together, these results suggest an important role for Vibrio in SSW, though future work in which the abundance of Vibrio is manipulated would be needed to test this hypothesis.

While Vibrio species may be sufficient in causing disease in some cases, other opportunistic taxa may also play an important pathogenic role, taking advantage of a weakened immune system and further contributing to the anaerobic and lethal nature of SSW. A number of the differentially abundant taxa identified in the present study are linked to disease in humans and animals. Microbes in the order Clostridiales are implicated in several human intestinal diseases due to the secretion of harmful toxins (Bauer and Kuijper, 2017). Some members of Fusobacteriaceae are known to cause skin infections and topical ulcers on their hosts through the secretion of harmful metabolites (Olsen, 2014). Additionally, Moritellaceae has been characterized as a fish pathogen with severe economic impact, forming lesions on their teleost hosts (Urakawa, 2014). The mechanistic role in disease progression of the anaerobes that increase with signs of SSW remains up for debate. It is likely that the suffocating properties of these proliferating anaerobes are not the sole driver of disease. Transcriptomic profiling of affected P. helianthoides unveiled a number of immune systems, tissue remodeling, and neural genes in response to SSW, which could be more than expected from anaerobic suffocation alone (Fuess et al., 2015).

We observed a gain and loss of taxa to a similar degree in each site-health comparison, yet generally only increases in pathway and KEGG functional category enrichment upon exposure and loss of pathways and KEGG enrichment with wasting signs ( Figure 2A ; Figure 4 ). Taken together, the initial increase of pathway functions may reflect a general colonization and/or proliferation of facultative anaerobes with exposure to SSW, while the low number of depleted pathways may reflect the loss of less-impactful taxa and/or an overlap in metabolic processes between those gained and lost. The transition to wasting signs is accompanied by the loss of pathway functions which may signify the loss of putatively beneficial microbes, and the lack of pathway enrichment with the onset of wasting may signify an overlap of pathway functions that were already increased by the initial SSW exposure. This suggests that only a few of the differentially expressed taxa may be responsible for the marked changes in pathway enrichment observed in this study, most likely related to the taxa with the largest effect size addressed above. However, 16S ribosomal RNA sequencing alone is not enough to fully characterize the metabolic roles of the colonizing taxa we have observed through the progression of SSW, and thus we are hesitant to make assumptions based on specific metabolic roles predicted in this analysis. Metagenomic and metabolomic analyses will be necessary to further characterize the molecular functions of the microbes and the potential role of secondary metabolites, secreted toxins, and virulence factors in the progression of this disease.

Polymicrobial diseases are difficult to study and thus-far poorly characterized in marine systems due to the variety of yet-unidentified microbes living in the oceans. However, many of these diseases share commonalities. Black band disease in corals, for example, is caused by the proliferation of cyanobacteria, sulfide-oxidizing, and sulfate-reducing bacteria on and around coral tissues similar to the taxa identified in this study (Sato et al., 2017). Other marine diseases, such as Pacific Oyster Mortality syndrome (PCOM), Bald Urchin Disease, and SKin Ulceration Disease (SKUD) are the result of opportunistic, and non-specific, bacterial colonization (Becker et al., 2008; Delroisse et al., 2020; Petton et al., 2021). PCOM was linked to a virally induced immunocompromised state acting as the catalyst for microbial dysbiosis and the opportunistic colonization by pathogenic bacteria eventually leading to animal death (Petton et al., 2021). Nonspecific viral infection such as Sea Star Associated DensoVirus (SSWaDV), along with climate anomalies, reduced dissolved oxygen, pollutants, and other biotic/abiotic stressors could all weaken the sea star’s immune response and promote the colonization of dysbiotic microbes. The impact of one or the cumulative effect of multiple stressful factors may tip the homeostasis of the microbial communities driving them toward dysbiosis and disease. While non-specific and opportunistic colonization is supported in many studies, our findings suggest certain microbial taxa may be more likely than others to initiate these early dysbiotic stages of SSW disease, acting as a steppingstone for more opportunistic pathogens to take hold.

The results of this study reveal a dysbiotic event preceding visible signs of wasting disease, driven by a few key taxa, and the persistence of these communities through the progression of the disease. The proliferation of specific microbes (most notably members of the genus Vibrio) prior to signs of disease implicates changes in biotic and abiotic conditions that may underlie the SSW epidemic, particularly in the case of microbial pathogens. In the future, extending approaches to better understand the host-microbe interactions at a higher resolution will allow investigators to better identify key microbes involved in both healthy and diseased states, their related metabolic functions, and subsequent consequences of microbiome disruption as a result of biotic and abiotic stress. Additionally, the function of the core microbiome in many natural systems remains unknown. Long-term monitoring of entire communities and their associated microbiomes may provide insights into the adaptive and evolutionary potential of the microbiome in the presence and persistence of the recent wasting outbreaks.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: BioProject, PRJNA931596.

All authors contributed to and approved the manuscript. Conception and design: ML, MP. Sample Collection: ML, KH. Data Generation: ML. Data analysis AM, BCh, DM, BCa, MP. Visualization: AM, BCh, DM, BCa. Writing – original draft: AM, BCh. Writing – review & editing: AM, BCh, DM, BCa, MP. All authors contributed to the article and approved the submitted version.