Introduction

Flavobacteria are considered to be important fish pathogens throughout the world. In particular, three species are reported to cause considerable losses in farmed and wild fish: Flavobacterium psychrophilum, Flavobacterium branchiophilum, and Flavobacterium columnare. F. psychrophilum is known as the causative agent of bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS) [1,2,3], F. branchiophilum causes bacterial gill disease [4,5,6], while F. columnare causes columnaris disease [4, 7]. There is a considerable number of further Flavobacteria species, which have been described as fish pathogens, e.g., F. johnsoniae [8], F. succinicans [9], and F. hydatis [10] (see also review by Loch and Faisal [11•] and Bernardet [12]). More recently, new species isolated from fish have been described, among them were F. chilense, F. araucananum [13], and F. spartansi [14]. In connection with an attempt to identify Flavobacterium species involved in disease events in Michigan, USA, it was suggested that there are a high number of species not described up to date [15]. Overall, F. psychrophilum and F. columnare are the two most investigated species affecting freshwater fish and causing major economic losses. For both, a widespread distribution is reported, with F. psychrophilum being linked to intensive salmonid aquaculture [1, 12] and F. columnare being the second most important pathogen in channel catfish farming [16]. Therefore, this review concentrates on the findings and research regarding these two species during the last 5 years.

Molecular Analysis of Genome

Great efforts have been invested to unravel the genetic background of both F. psychrophilum and F. columnare, where full genomes as well as parts thereof have been analyzed by different techniques.

F. psychrophilum

Multilocus sequence typing (MLST) offers an important technique to evaluate the genetic differences between global F. psychrophilum isolates. Respective investigations have been performed on a number of isolates from single European countries, e.g., France [17] and Switzerland [18] as well as from European geographic regions such as the northern European countries [19•], South America [20, 21], North America [22•], and Japan [23]. Some of the clonal complexes (CC) were found worldwide while others were restricted to a specific geographic origin. Moreover, in North America, a clear link between CC and fish species was suggested. Further, investigations of these North American isolates revealed that the majority of identified sequence types (ST) infected a single host species [24]. Collectively, these investigations identified a high diversity of isolates. A similar diversity was also discovered by other genotyping and serotyping techniques within isolates from UK [25].

Another approach to better understand the genetic background of mechanisms involved in virulence, transmission, pathology, serotypes, and protein expression is full genome sequencing. Such analyses have been performed on different isolates from various species and geographic origins in recent years [26,27,28,29,30]. Again, a wide variety within the genome was found between different isolates and a clear host specificity of different isolates was evident [28]. In addition, by complete genome sequence analyses and serotyping methods, an association between PCR-serotype and host species was demonstrated [29].

F. columnare

Likewise, different techniques to disentangle the genetic composition of F. columnare have been applied. Based on partial sequencing of the 16S rRNA gene of all known genomovars, a standard protocol to type isolates of F. columnare by restriction fragment length polymorphism (RFLP) was developed [31]. The technique was optimized by designing a new degenerate reverse primer for the amplifying of a 1250-bp product of the 16S rRNA gene [32•]. Amplification of the almost full 16S rRNA gene and subsequent RFPL techniques demonstrated that in the Great Lakes Basin (North America), infection of fish appears to be caused by isolates belonging to one particular genomovar, i.e., genomovar I [33]. This was concluded from the analyses of a subset of 34 out of over 300 isolates from a variety of different fish species. In contrast, in Brazil, F. columnare isolates retrieved from the same fish species were found to belong to different genomovars [34]. By means of repetitive extragenic palindromic PCR (REP-PCR), the same authors also demonstrated considerable intragenomovar diversity. A similar diversity of isolates belonging to the same genomovar and the same geographic region (Idaho, USA) was recorded based on amplified fragment length polymorphism fingerprinting (AFLP) [35].

Full gene sequencing has been performed for F. columnare isolates from a variety of fish species and belonging to different genomovars [36,37,38,39,40,41]. Isolates belonging to different genomovars revealed clear genetic differences [42•] but functional similarities [39]. The degree of diversity between some isolates tempted researchers to reconsider the actual affiliation of some isolates and to suggest that some isolates might belong to a novel taxa or subtaxa within the genus Flavobacterium [42•]. Sequence analyses allowed the identification of regions encoding for proteins and to predict their functions [43], although the functions of many proteins remain still to be determined [44]. Comparison of full sequences of different genomovars and other Flavobacterium species suggested that F. columnare genomovars I and II are more closely related to each other than to F. psychrophilum or F. johnsoniae [44].

Genetic Aspects and Pathogenicity/Host Specificity

Unraveling the genetic background of a pathogen allows a better understanding and elucidation of the relationship between different isolates as well as mechanisms involved in pathogenicity. For both species, major advances in this field have been achieved in the last 5 years.

F. psychrophilum

The pan-genome of F. psychrophilum was described based on 11 isolates from temporally and spatially different populations (Chile, North America, Denmark, and France) [45]. As with every new genome additional genes were detected, the authors concluded that F. psychrophilum possesses an open pan-genome [45], indicating that with all the isolates sequenced so far, the diversity is still far from being covered. Putative virulence factors were evenly distributed among the isolates matching with phenotypic traits such as biofilm production and secretion of extracellular enzymes. From these findings, the authors concluded that F. psychrophilum isolates have similar modes of adhesion, colonization, and pathogenicity across large spatial and temporal spaces [45].

F. columnare

Full genome sequencing of a F. columnare isolate allowed to determine a range of potential virulence factors associated with different modes of action as well as genes responsible for a variety of secreted proteases [46•]. Li et al. [47] demonstrated the potential involvement of a type IX secretion system (T9SS) in the secretion of extracellular proteins and molecules associated with the gliding motility. Mutants lacking a core component of T9SS had lost the gliding motility and were clearly less virulent to zebrafish (Danio rerio), channel catfish (Ictalurus punctatus), and rainbow trout (Oncorhynchus mykiss) [47]. A relation between colony type and expression of chondroitinase encoded by cslA was established [48], as the virulent phenotype expressed significantly more of this proteolytic enzyme. The importance of chondroitinase as a virulence factor was also been shown by Li et al. [49, 50] who found a significantly lower pathogenicity in mutants lacking the sequence for encoding chondroitinase. Zhang et al. [51] produced mutants lacking two genes encoding for proteins with a conserved glycoside hydrolase family domain. Single and double-gene mutants still formed rhizoid and non-rhizoid morphotypes. However, comparison of the gene expression patterns between the two morphotypes revealed higher expression of several genes in the rhizoid forms suggesting that these genes might be related to the higher virulence found in zebrafish experiments [51].

Genetic Relationship of Isolates and Geographical Distribution of Variants

F. psychrophilum

The geographic range, in which F. psychrophilum is recorded, constantly expands. As an example, the pathogen has been described for the first time in 2016 from cultured rainbow trout in Argentina [52] and Mexico [53]. However, it is not really clear whether this is a true spread of the bacterium or more a question of first detection.

F. columnare

As described for F. psychrophilum, the geographic range of recorded F. columnare isolates is also expanding. In 2013, the pathogen was recorded for the first time in farmed Catla catla in India [54]. Moreover, within areas of Thailand, isolates recovered from red tilapia demonstrated to have similar phenotypic characteristics but a clear genetic variety [55]. All the analyzed isolates but one belonged to genomovar II, indicating co-existence of different genomovars within the country. Interestingly, a subset of the isolates also contained a smaller intergenic spacer region (ISR) thus, forming a unique phylogenetic group, so far only described from Thailand [55].

Development of Vaccines

With Flavobacteria infections being a major problem in many aquaculture facilities, effective prophylactic and treatment methods are of vital importance. Vaccination has shown to be an appropriate measure for a variety of bacterial diseases. However, so far, attempts against both F. psychrophilum and F. columnare have not yet yielded full protection.

F. psychrophilum

Optimization of the vaccination effect using an attenuated immersion procedure was achieved, when fish were first boostered after 12 weeks initial vaccination [56]. In this report, a specifically rifampicin-resistant strain had been produced by passages against increasing concentrations of rifampicin, which showed a high degree of attenuation. Although a second booster did not result in higher relative percent survival (RPS) values, nevertheless RPS was increased by a longer exposure to the vaccine immersion (30 instead of 3 min). Efficacy of vaccination was further refined by feeding a “health enhancing” diet to fish [57]. Hoare et al. [58] could significantly increase RPS in rainbow trout immersion vaccinated with a polyvalent whole cell vaccine containing formalin inactivated F. psychrophilum. The authors found increased levels of immunoglobulin T (IgT) in systemic organs. However, it was not clear if IgT played a functional role in protection. In this context, both oral and i.p. vaccination did not significantly differ using a live-attenuated vaccine based on a rifampicin-resistant F. psychrophilum strain, neither was there a significant difference, when bacterial cells were given microencapsulated or unencapsulated [59]. Nakayama et al. [60] explored the potential of F. psychrophilum collagenase as an antigen in an immersion vaccine for ayu (Plecoglossus altivelis). Recombinant F. psychrophilum collagenase was expressed in Escherichia coli and Brevibacillus chosinensis, and the culture supernatant of B. chosinensis was used as an immersion vaccine solution. Vaccinated ayu was bath challenged with F. psychrophilum, and the results revealed a protective effect in challenged ayu [60]. Kato et al. [61] described three surface proteins (3-hydroxyacyl-CoA dehydrogenase (HCD), ATP synthase beta subunit (atpD), and glutamate dehydrogenase (gdhA)) as being protective antigens against F. psychrophilum which were therefore considered as good candidates for development of vaccines against this pathogen in ayu. Plant et al. [62] validated a recombinant F. psychrophilum gliding motility protein as a vaccine candidate without success. This is in agreement with Gomez et al. [63•] who detailed the difficulties in developing an effective commercial F. psychrophilum vaccine.

F. columnare

While live vaccines always represent a potential danger of reversion, use of ghost cells does exclude this potential, as the cells used have expelled the cytoplasmic cell content but still have the complete envelope representing the antigenic pattern. Use of F. columnare ghost cells as a vaccine resulted in 70% RPS compared to 41.9% with a vaccine using formalin killed bacteria in intraperitoneally vaccinated grass carp (Ctenopharyngodon idellus), thus providing a promising alternative to full pathogen-based vaccines [64]. Further attempts to produce an effective vaccine were performed by searching for immunogenically active bacterial proteins [51, 65,66,67]. In challenge experiments, vaccines based on identified proteins (gliding motility lipoprotein GldJ (GldJ), lipoprotein (Lip), and outer membrane efflux protein precursor (Omep)) [65] or a recombination of different proteins (gliding motility lipoprotein GldJ (GldJ), hypothetical protein FCOL_13420 (Fco1), lipoprotein (Lip), F0F1 ATP synthase subunit beta (F0f1), and outer membrane efflux protein precursor (Omep)) [51] yielded some degree of protection against F. columnare.

Development of Alternative Treatment and Prophylactic Methods: Probiotics, Plant Extracts, Phages

Development of resistance to a wide variety of commonly used antimicrobials and lack of efficient and registered vaccines make alternative methods to treat flavobacterial infections particularly desirable. A number of investigations in this field have been performed in the last 5 years.

F. psychrophilum

Disinfection of eggs in the pre-water hardening phase (fertilized eggs before the chorion becomes rigid due to exposure to the water) with povidone iodine (50 ppm) in salmonid hatcheries in Japan proved to be a promising method for prevention of F. psychrophilum infection of eggs [68]. Supplementation of extenders used for disinfection of rainbow trout sperm with penicillin and streptomycin was successful in reducing subsequent F. psychrophilum infection, but there were some factors like lower egg fertilization and survival that needed further investigation [69]. The use of the same antibiotics added to a 0.5% NaCl fertilization diluent and hatchery well water during water hardening significantly reduced F. psychrophilum on the surface and inside eggs compared to controls [70]. Also, a combination of the two antibiotics and elevated temperatures killed the bacterium under laboratory conditions [71]. However, there might be a discrepancy between in vitro results and in vivo effectivity as shown for enrofloxacin, doxycycline, and florfenicol when treating RTFS-affected trout [72•].

In treatment experiments, copper sulfate for the control of F. psychrophilum infections was not found to be suitable as the concentrations required to eliminate the bacteria were toxic to the eggs [73].

Long et al. [74] suggested broodstock screening as a potential tool for evaluating F. psychrophilum infection levels, which could become an important component for disease management.

Different substances and techniques related to the diet were tested for their preventing effect on infections by F. psychrophilum. Trout by-product hydrolysates, generated using trout pepsin, had a high antibacterial activity against F. psychrophilum [75]. Rainbow trout fed a diet naturally contaminated with deoxynivalenol were better protected against an experimental F. psychrophilum infection compared to controls [76]. Also, restricted feed intake was shown to have a protective effect for rainbow trout infected with the pathogen [76]. The effect of a vaccine could be improved, when fish were fed a “health enhancing” diet [57]. Administration (either orally or intraperitoneally) of a probiotic Enterobacter species was a potential method to protect fry from disease outbreaks with F. psychrophilum, [59, 77]. In contrast, prebiotic supplementation of diets with 30% soy bean meal only had a minimal effect on disease resistance of cutthroat trout (Oncorhynchus clarkii lewisi) experimentally infected with F. psychrophilum by a cohabitation method [78]. On the other hand, probiotic treatment of the tank water decreased the population of F. psychrophilum in the water [79].

Phage-based treatment of F. psychrophilum was suggested [80•] and showed promising features in laboratory experiments [81•]. Phage resistance was seen but the small populations of resistant clones of bacteria had reduced competitive abilities relative to the sensitive ancestral strain [81•, 82]. Phages were shown to be able to survive passage through the fish stomach, to penetrate the intestinal barrier, and to enter the circulatory system after oral delivery, meaning that delivery of phages via coated feed pellets could constitute a potential way of treatment as well as prevention of disease with F. psychrophilum [83].

F. columnare

Also, for prevention of F. columnare infections, a variety of methods and substances were evaluated. Thus, chloramine-T and hydrogen peroxide treatment were effective in controlling mortality associated with external columnaris in bass and bluegills [84]. A positive effect regarding survival to columnaris disease was obtained in several fish species by addition of extra copper sulfate to the feed [85,86,87]. Kaolinitic clay added to the water was found to have a positive effect on the survival of channel catfish experimentally infected with F. columnare [88]. Channel catfish kept in water with kaolin were found to have less gill damage than control fish as it seemed that kaolin reduced the adhesion of the bacteria to the gills [88]. Straus et al. [89] found that the concentration of divalent cations (hardness) in the water affected the pathogenesis of columnaris disease. On the other hand, environmental nutrients may act as significant triggers of virulence gene expression, in this case, in F. columnare [90•].

Chitosan showed antibacterial activities and was a promising candidate to be explored as alternative to antibiotics [91]. Nigella sativa (black cumin) seeds and oil have been found to inhibit the growth of F. columnare strains in the laboratory as well as to have a positive effect on survival of catfish and zebrafish when added to their diets [92]. Grass carp fed with diets supplied with Ficus carica polysaccharide were more resistant against F. columnare infections [93]. Also, ungeremine and analogues were successfully tested as bactericides against F. columnare [94]. Wogonin (a flavone) was effective in vitro against F. columnare [95]. Deficiency in vitamin C in the diet depressed the gill physical and immune barriers in grass carp under infection with F. columnare, while at optimal dietary levels, the barrier function was improved [96].

Different prebiotic dietary additives reduced the susceptibility for columnaris infection [97]. Two Pseudomonas fluorescens strains isolated from the skin and gut of healthy walleye (Sander vitreus) and used as probiotics had a positive effect with an improved survival when it comes to F. columnare disease in walleye [98].

Phages are an alternative to chemotherapy against columnaris disease in aquaculture as demonstrated in controlled experiments by Laanto et al. [99].

Investigations of Different Susceptibilities of Fish Strains and Species to the Pathogens

F. psychrophilum

Within the same fish species, rainbow trout, differences in virulence of F. psychrophilum were found depending on the fish strain used [100]. A possible explanation was derived from a histological study, as differences in severity but not in quality of the disease were found between resistant and susceptible strains of rainbow trout [101]. Upon challenge with F. psychrophilum, susceptible and resistant strains of rainbow trout were characterized by different expression of genes encoding for proteins particularly involved in the innate immune response [102]. An investigation of isolates retrieved from different Oncorhynchus species from the Great Lake Basin in North America revealed a genetically heterogenic F. psychrophilum population and an association between sequence variation and fish host species [22•]. Thus, some bacterial strains were specific to either rainbow trout, Chinook salmon (Oncorhynchus tshawytscha), or coho salmon (Oncorhynchus kisutch), while others did not display a marked host specificity. A differing virulence for different salmonid species was also found for Chilean F. psychrophilum isolates, where isolates highly pathogenic for rainbow trout did not induce mortality in Atlantic salmon (Salmo salar) [103]. When investigating isolates from Japan belonging to four genotypes retrieved from four different fish families, only one genotype displayed a strong pathogenicity for ayu (Plecoglossus altivelis), while the other three genotypes had none at all or only a weak pathogenicity [104]. In a further study, captive and wild salmonids were investigated for infection with F. psychrophilum, and it was shown that wild fish showed a clearly greater prevalence of the bacterium compared to captive fish [105].

F. columnare

Channel catfish (Ictalurus punctatus) families specifically bred for Edwardsiella ictaluri resistance proved to be more susceptible to F. columnare than families which were less resistant to E. ictaluri [106]. This demonstrates that there is not necessarily a common background of resistance to bacterial pathogens. Comparison of hybrid striped bass and white bass (Morone chrysops) in challenge experiments with F. columnare revealed a higher susceptibility of the hybrid strain expressed by a higher degree of gill lesions and lower survival rate [107]. Analyses of gene expression in gills of catfish demonstrated clearly different patterns of expression mainly of genes involved in innate immune response and mucus secretion between susceptible and resistant breeds [108].

Bacterium Detection and Growth Media

Important prerequisites to handle Flavobacterium infections are sensitive and reliable detection methods as well as suitable culture media to test isolates for growth conditions and sensitivity to antibiotics.

F. psychrophilum

Different approaches to identify F. psychrophilum have been developed, among them matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS) [109]. According to Fernández-Álvarez et al. [110], this method allowed to discriminate between serotypes, whereas in other laboratories, this was not possible (Madsen, personal communication). Determination of serogroups is also possible by a newly developed multiplex PCR [29]. Advances have also been made with respect to the detection of F. psychrophilum in river water in order to get an idea on the environmental distribution of this species [111]. In a series of experiments, Long et al. [74] applied ELISA and PCR to detect F. psychrophilum in organs, ovarian fluid, and progeny of rainbow trout and coho salmon to investigate the rate of potential vertical transmission. Interestingly, determination of the pathogen in ovarian fluid was not a reliable method while a good correlation between bacteria in the kidney of broodstock (measured by ELISA) and infection level of progeny could be established.

When isolating bacteria from fish or the environment, the suitability of the medium plays a crucial role. In this respect, Oplinger and Wagner [112] compared different culture media, i.e., standard enriched Anacker-Ordal broth (EAO) with different supplementations, tryptone-yeast extract salts, and maltose infused tryptone-yeast extract salts. They found the best growth in EAO, whereby supplementations did not significantly improve the result. The applied method is also important when it comes to resistance testing as exemplified by Gieseker et al. [113] and Miranda et al. [114] who standardized and optimized methods to determine susceptibility of isolates to antibiotics. Miranda et al. [114] considered MIC determination as the preferable method for antimicrobial susceptibility testing of F. psychrophilum in comparison with disk diffusion, due to the more precise data for the former compared to the latter.

F. columnare

A comparison of different detection methods revealed species-specific PCR to be more sensitive than conventional culture methods [115]. Several adaptations and improvements of detection methods have been effected. A single tube colorimetric method of loop-mediated isothermal amplification reliably detected F. columnare in different organs of apparently healthy tilapia with a ten times higher sensitivity than conventional PCR [116]. By optimizing the reverse primer of a qPCR method, it became possible to amplify also the 16S rRNA of isolates from Thailand, which was not possible before [32•]. Gibbs [117] developed and validated a qPCR which recognizes F. columnare regardless of the genomovar both in fish tissue and pond water, respectively.

Investigations on suitable culture media for susceptibility testing to antibiotics have also been performed for F. columnare. Gao and Gaunt [118] compared the growth of different F. columnare isolates on three broths and three agars. Further, different ingredients were added or subtracted to the media and the effect on bacterial growth was tested. Most rapid and uniform growth was obtained with a newly developed G medium. Gieseker et al. [113] performed a multi-laboratory trial to test a broth microdilution method and later further improved this method [119].

Biology and Disease

New insights in mechanisms of pathogenicity and transmission are important to understand the disease and influencing factors. New information on this topic has been published mainly for F. columnare.

F. psychrophilum

Attempts were made to unravel mechanisms of pathogenesis. In an in vitro experiment, Iturriaga et al. [120] could demonstrate that F. psychrophilum induces apoptosis in muscle cells through the modulation of the NF-κB signaling pathway. In exposure experiments, F. psychrophilum cells of smooth and rough colonies showed a tissue tropism with the highest degree for fin tissue, independent of the colony type [121]. However, bacteria from the smooth colony type remained only shortly on the host tissue. Skin mucus promoted growth of both colony phenotypes [121].

Vertical transmission is still under debate. An investigation of eggs from Atlantic salmon held as captive broodfish showed the predominant bacterial species on the surface to be Pseudomonas fluorescens. However, from within the eggs, different concentrations of F. psychrophilum depending on the origin of fish could be isolated [122] indicative for an intra-ovum infection, although it has to be kept in mind that bacteria had been isolated from dead eggs only.

F. columnare

Biofilm formation is an important parameter for the pathogenicity of a bacterium, and its formation on biotic (host tissue) and abiotic surfaces decreases the effectiveness of antimicrobials and the host defense mechanisms. Extracellular products (EPS) contribute to biofilm formation Hence, knowledge on the composition might help to find measures against biofilm production. An investigation of EPS of F. columnare isolates from Brazil revealed that the predominant monosaccharide is glucose [123]. Enhancement of biofilm production has been achieved by providing L-rhamnose [124]. Bacteria grown at higher concentrations of L-rhamnose increased mortality in catfish, further demonstrating the importance of biofilm production as a virulence factor. Another factor suggested to be involved in virulence is the potential of iron uptake. A siderophore-mediated iron uptake system has been demonstrated in F. columnare [125]. The importance of iron uptake in virulence was corroborated in a comparison of a highly virulent and a low virulent F. columnare strain, with the former having manifold higher expression of genes involved in iron uptake [126]. Chondroitin AC lyase has also shown to function as a virulence factor. The genetic background of this factor has been investigated by producing mutants devoid of a gene involved in expression of the enzyme [127]. Virulence of an enzyme-deficient mutant was fourfold lower in a zebrafish model [128]. The role of chondroitinase as a virulence factor has further been confirmed by its higher expression in the virulent (rough) colony type [48]. Virulence is also influenced by the amount of mucus shed by fish, as mucus has shown to support growth and survival of F. columnare [129]. Kinnula [130] found a relation between infection dose, nutrient load in the water, co-infection with different isolates, and virulence in zebrafish and rainbow trout. The role of nutrient load in the water has also been shown by Penttinen [48]. Additionally, water hardness seems to influence virulence of F. columnare [89]. As already mentioned, the morphotype of colonies plays an important role in regard to virulence, whereby the rhizoid type is more virulent than non-rhizoid colonies [131, 132]. Virulence might also be a result of a selection process as revealed by Kunttu et al. [133] who demonstrated that isolates from fish farms were more virulent than environmental isolates not influenced by fish farming activities. This hypothesis is partly supported by the finding of a non-existing sequence and very low protein variation among 70 highly virulent genomovar I isolates from one fish farm in Idaho [134].

Testing of isolates for virulence and pathogenesis requires suitable challenge models. LaFrentz et al. [135] presented a challenge model for waterborne infection of rainbow trout, and Declercq et al. [136] developed challenge models for carp (Cyprinus carpio) and rainbow trout.

Further factors influencing the outcome of columnaris disease are co-infections with other pathogens. In both goldfish (Carassius auratus) infected with Dactylogyrus intermedius [137] and in hybrid tilapia (Oreochromis spp.) infected with Ichthyophthirius multifiliis [138], susceptibility for F. columnare was enhanced.

F. columnare is regularly demonstrated in new organs or new fish species. As an example, F. columnare-like bacteria were isolated from the brain of Nile tilapia (Oreochromis niloticus) from three fish farms in Thailand [139]. Major fish kills in the environment were attributed to F. columnare, e.g., in bluegill (Lepomis macrochirus), an infection with genomovar II was described for the first time [140]. In southern Saskatchewan lakes in Canada, die-offs of yellow perch (Perca flavescens) and lake whitefish (Coregonus clupeaformis) were related to an infection by F. columnare and Aeromonas hydrophila [141].

Conclusion

Research within both F. psychrophilum and F. columnare produces new insights in virulence factors, host susceptibility, genetic background, and treatment methods. When comparing the two species, it becomes evident that there is considerable agreement with regard to the problems caused, the topics investigated, and the avenues to be explored for future research. Despite all the efforts to disclose pathogen mechanisms and to find suitable methods to prevent or treat infection, the two bacteria still represent a major threat to aquaculture, both in temperate and warm climate zones, and effective, internationally registered vaccines are still lacking. Furthermore, reported die-offs from wild fish populations illustrate the importance of these pathogens.