**3. Discussion**

The studies identified in this review have demonstrated the failure of many common disinfection protocols to achieve long term elimination of *L. pneumophila* from hospital and potable water supplies when protozoan hosts are present [35,38] (as mentioned in Table 1). This long term survival could be attributed to association with biofilms, inherent tolerance of *L. pneumophila* to high temperature and chemical disinfectants, and constant reseeding from source water [59]. However, perhaps the most interesting and undervalued relationship is the interactions with protozoan hosts. The studies identified (Table 1) are from 14 different countries, which demonstrates the need for further research to understand the *L. pneumophila*–protozoan interaction under different environmental conditions found globally. Proper managemen<sup>t</sup> of legionellosis requires a better understanding of *L. pneumophila–*protozoan interaction, the diversity of protozoan hosts in hospital and potable water systems and the role of the host in bacterial survival under different environmental conditions.

#### *3.1. Implications for the Control of L. pneumophila*

Numerous studies have demonstrated the presence of *L. pneumophila* in disinfected water supplies [60,61]. An important factor enabling *L. pneumophila* survival in the built environment is its interaction with a protozoan host [62–64] (as mentioned in Table 3). Thermal treatment is one of the most common methods used to disinfect hospitals and building water supplies. In the USA [35], Germany [38] and Slovakia [41], thermal disinfection was adopted for managemen<sup>t</sup> of nosocomial

outbreaks of legionellosis. This strategy was unable to maintain water control for a long period of time [35,38] (as mentioned in Table 1). Rhoads et al. [64] reported that *L. pneumophila* associated with *V. vermiformis* can tolerate thermal (58 ◦C) treatment, and this disinfection protocol is unable to reduce microbial load in water. Published evidence suggests *Legionella* associated with *Acanthamoeba* are more thermos-tolerant and can survive at even higher temperatures ranging from 68–93 ◦C [63]. According to Steinert et al. [38] members of *L. pneumophila* SG1 are more thermo-tolerant than SG2. This is significant given the high number of legionellosis cases associated with *L. pneumophila* SG1.

As per WHO guidelines [65], 0.2 mg/<sup>L</sup> of free residual chlorine at point of delivery is recommended in potable water for disinfection. A pilot scale study conducted by Muraca et al. [66] demonstrated that 4 to 6 mg/<sup>L</sup> chlorine treatment for 6 h resulted in 5–6 log reduction of *L. pneumophila*. It was also observed that the e fficacy of chlorine against *Legionella* was enhanced at 43 ◦C. However, at high temperatures a continuous flow of chlorine was required to overcome thermal decomposition. In vitro studies demonstrated higher level of tolerance to free chlorine (up to >50 mg/L) when bacteria are associated with host *Acanthamoeba* cysts [67]. According to Kool et al. [68], water disinfection with monochloramine resulted in a reduction of nosocomial LD outbreaks in USA. However, other studies have shown that some strains of *L. pneumophila* can tolerate high levels of monochloramine disinfection (17 mg-min/<sup>L</sup> for 3 log reduction) [69]. Donlan et al. [70] reported that *L. pneumophila* associated with amoebae in biofilm are less susceptible to chlorine and monochloramine treatment. It is also reported that monochloramine disinfection in hospital settings results in transformation of *L. pneumophila* vegetative cells to VBNC state [27].

According to Walker et al. [71] chlorine dioxide can e ffectively control *L. pneumophila* from hospital water system. In vitro studies demonstrated that 0.4 mg-min/<sup>L</sup> residual chlorine dioxide achieved a 3 log reduction of *L. pneumophila*. However, this procedure was not e ffective for amoebae associated *L. pneumophila* [69]. According to Schwartz et al. [72] *Legionella* biofilms on polyvinyl chloride, polyethylene and stainless steel materials can tolerate chlorine dioxide treatment. Muraca et al. [66] conducted a pilot scale study and reported that 1–2 mg/<sup>L</sup> residual concentration O3 treatment for 5 h resulted in 5 log reduction of *L. pneumophila*. However, half-life of ozone in water is very short, so it is di fficult to maintain residual concentration in water supplies. According to Wang et al. [54], if chlorination and ozonisation is used in combination, it can target both *L. pneumophila* and its host protozoans e ffectively. In combination both treatments e ffectively eliminated planktonic *L. pneumophila* and free living *Naegleria* from water, whereas this combination could only reduce the population of *Acanthamoeba* (≈0.9 log10 gene copies/g). In comparison to chlorination alone, this combination method significantly reduced the population of *L. pneumophila* (≈3 log10 gene copies/g) and host amoebae (≈3 log10 *Naegleria* gene copies/g and ≈6.1 log10 *Acanthamoeba* gene copies/g) co-existing in biofilms.

UV irradiation is another method of disinfection. These radiations harbor strong genotoxic attributes. Cervero-Arago et al. [73] demonstrated that 5–6 mJ/cm<sup>2</sup> UV dose was su fficient to achieve 4 log reduction *L. pneumophila* population. According to Muraca et al. [66] 30 mJ/cm<sup>2</sup> UV rays treatment for 20 min resulted in 5 log reduction of *L. pneumophila*. However, continued exposure to same fluence rate for 6 h unable to eliminate all culturable *L. pneumophila* (1–2 × 10<sup>2</sup> CFU/mL). Schwartz et al. [72] reported that *Legionella* biofilms on stainless steel, polyvinyl chloride and polyethylene surfaces can tolerate UV treatment. It was also reported that amoebae associated *L. pneumophila* can tolerate much higher doses of UV rays [73].

#### *3.2. Protozoan Host Control Strategies*

Protozoans present in water supplies play an important role in *L. pneumophila* survival and resistance against disinfection protocols. Interesting, it has also been suggested that some protozoans infected by *L. pneumophila* have increased resistance to disinfection procedures compared to those uninfected [74]. As such, an understanding of protozoan disinfectant resistance and *L.pneuophila–*protozoan interactions is essential for the improved managemen<sup>t</sup> of manufactured water systems. According to Loret et al. [75], common water chemical disinfection protocols, i.e., ozonisation (0.5 mg/L), chlorination (free chlorine 2 mg/L), electro-chlorination (free chlorine 2 mg/L), monochloramine (free chlorine 2 mg/L), chlorine dioxide (0.5 mg/L) and Cu+/Ag+ ions (0.5/0.001 mg/L) treatments, are unable to completely eliminate amoebae cysts hosting *Legionella* from water supplies (Table 3). These methods appear to be only effective against the free living amoebae population, as they are not feasible for targeting biofilm-associated amoebae [76]. The non-standardized approach to evaluating disinfection limits is one of the gaps in knowledge raised in the discussion section.

In vitro studies have shown 1 mg/<sup>L</sup> chlorine is sufficient to inhibit the growth of *Acanthamoeba*, *Vermamoeba* and *Vahlkampfia* trophozoites. Importantly, after two hours exposure, chlorine produced complete die-off of trophozoites [77]. According to Kuchta et al. [78] 2–4 mg/<sup>L</sup> chlorine treatment for 30 min can completely inactivate *Vermamoeba* trophozoites. Whereas, trophozoites of some strains of *Hartmannella* required 15 mg-min/<sup>L</sup> chlorine treatment for only 2 log reduction [79]. Mogoa et al. [80] reported that *Acanthamoeba* trophozoites exposed to 5 mg/<sup>L</sup> chlorine for 30 s resulted in a 3 log population reduction. It was also demonstrated that in *Acanthamoeba*, chlorination induces various cellular changes including reduction in cell size and alterations in cellular permeability. Dupuy et al. [79] noticed that *Acanthamoeba* trophozoites treated with 28 mg/<sup>L</sup> chlorine for 1 min only resulted in a 2 log reduction. In comparison with uninfected *Acanthamoeba* trophozoites, *L. pneumophila* infected *Acanthamoeba* trophozoites were more resistant against sodium hypochlorite (1024 mg/L) treatment [74].

Generally, inactivation of *Acanthamoeba* and *Vermamoeba* cysts required 5 mg/<sup>L</sup> chlorine, whereas for *Vahlkampfia* 2 mg/<sup>L</sup> chlorine treatment. It is important to note that cysts of *Acanthamoeba* were found highly resistant and only a 2 log reduction was noticed after eight hours exposure [77]. It was also reported that *Acanthamoeba* cysts can tolerate 100 mg/<sup>L</sup> of chlorine for 10 min [81]. According to Dupuy et al. [79] treatment of *Acanthamoeba* cysts with 856 mg-min/<sup>L</sup> results in only 2 log reduction. Loret et al. [82] reported that to achieve 4 log reduction for *Acanthamoeba polyphaga* cysts 3500 mg-min/<sup>L</sup> chlorine treatment is required. Likewise certain strains of *Hartmannella* cysts can tolerate high dose of chlorine (2 log reduction by 156 mg-min/L) [79]. Exposure of *Vermamoeba* cysts to 15 mg/<sup>L</sup> of chlorine for 10 min was lethal and resulted in complete inactivation [83].

Unlike *Acanthamoeba* and *Vermamoeba*, trophozoites and cysts of *Naegleria* were found sensitive to available disinfection protocols. *Naegleria* trophozoites were sensitive to 0.79 mg/<sup>L</sup> chlorine treatment for 30 min [84], whereas cysts were inactivated after exposure to 1.5 mg/<sup>L</sup> chlorine for 1 h [85]. Dupuy et al. [79] reported that chlorine treatment of *Naegleria* trophozoites with 5 mg-min/<sup>L</sup> resulted in only 2 log reduction and cysts can tolerate much higher levels of chlorine (29 mg-min/<sup>L</sup> for 2 log reduction). In potable water *Naegleria fowleri* associated with biofilms was able to tolerate 20 mg/<sup>L</sup> chlorine for 3 h [86].

In comparison to chlorine, chloramine is regarded as more stable disinfectant and capable to penetrate complex biofilms [68]. Dupuy et al. [79] suggested that instead of chlorine, monochloramine is effective chemical disinfectants against trophozoites and cysts of *Acanthamoeba*, *Vermamoeba* and *Naegleria*. It is possible that monochloramine harbors greater penetrating power than chlorine and easily enter in trophozoites and cysts. According to Mogoa et al. [87] monochloramine specifically targets the cell surface of *Acanthamoeba*. Dupuy et al. [79] identified that 352 mg-min/<sup>L</sup> monochloramine exposure resulted in 2 log reduction of *Acanthamoeba* cysts. Goudot et al. [88] noticed that 4–17 mg/<sup>L</sup> monochloramine exposure for 1 min only resulted in 2 log reduction of both planktonic and biofilm associated *Naegleria*. According to Dupuy et al. [79] to achieve 2 log reduction of *Hartmannella* trophozoites and cysts 12 mg-min/<sup>L</sup> and 34 mg-min/<sup>L</sup> monochloramine dose is required, respectively. Although in vitro studies demonstrate that higher concentration of chlorine-based disinfectants can inhibit the proliferation of protozoans; however, it can corrode the plumbing system pipes.

Chlorine dioxide has been shown to easily penetrate into amoeba trophozoites and cysts and specifically promotes cytoplasmic vacuolization in *Acanthamoeba* [87]. However the efficacy of chlorine dioxide varies from amoeba strains. The cys<sup>t</sup> form of some *Acanthamoeba* strains have been demonstrated to be highly tolerant to chlorine dioxide (35 mg-min/<sup>L</sup> for 2 log reduction) [79]. Loret et al. [82] demonstrated that an 80 mg-min/<sup>L</sup> dose of chlorine dioxide is required to achieve 4 log reduction of *Acanthamoeba polyphaga* cysts. Importantly, most studies were designed to investigate the effect of disinfection procedures on amoeba and there are limited studies on *L. pneumophila*-amoebae interactions during disinfection.

Ozonisation is an effective method of water disinfection. According to Cursons et al. [84], a dose of ozone 6.75 mg/<sup>L</sup> (0.08 mg/<sup>L</sup> residual level after 30 min) was sufficient to kill 99.9% (3 log reduction) trophozoites of *Acanthamoeba* and *Naegleria*. However, biofilm associated *Acanthamoeba*, *Hartmannella*, and *Vahlkampfia* were always found resistant to such treatments [76]. Loret et al. [82] demonstrated that 10 mg-min/<sup>L</sup> ozone dose resulted in 3 log reduction of *Acanthamoeba* trophozoites, however cysts retained viability.

Thermal treatment is a common physical disinfection protocol used for potable water supplies. According to Chang [89] trophozoites of *Naegleria* can survive at 55 ◦C for 15 min, whereas cysts can tolerate 65 ◦C for 3 min. *Vermamoeba* trophozoites and cysts have been shown to be completely inactivated by exposure to 60 ◦C for 30 min [78,83]. Thermal treatment of *Acanthamoeba* trophozoites and cysts at 65 ◦C for 10 min resulted in full inactivation [90]. Loret et al. [82] demonstrated that thermal treatment of *Acanthamoeba polyphaga* cysts at 65 ◦C for 120 min resulted in 5 log reduction. However, Storey et al. [81] reported that *Acanthamoeba castellanii* cysts are thermally stable and retain viability at 80 ◦C for 10 min. It has also been reported that thermal treatment can enhance the efficiency of chlorination. Although at high temperature (50 ◦C) the solubility of chlorine gas in water decreases significantly and very corrosive to pipe work, but its amoebicidal activity increases slightly [69].

UV treatment is another method of disinfection recommended by WHO. As per recommendation in 10 mJ/cm<sup>2</sup> dose is sufficient for 99.9% (3 log) inactivation of protozoans like *Giardia* and *Cryptosporidium* [65]. According to Cervero-Arago et al. [73] to achieve 3 log reduction of *V. vermiformis* trophozoites 26 mJ/cm<sup>2</sup> UV dose was required, whereas 76.2 mJ/cm<sup>2</sup> for cysts. It was also noticed that exposure to 72.2 mJ/cm<sup>2</sup> irradiance resulted in 3 log reduction of *Acanthamoeba* trophozoites [73]. Aksozek et al. [91] reported viability of *Acanthamoeba castellanii* cysts after exposure to high doses of UV rays (800 mJ/cm2). According to Sarkar and Gerba [92] to achieve 4 log reduction of *Naegleria fowleri* trophozoites and cysts 24 mJ/cm<sup>2</sup> and 121 mJ/cm<sup>2</sup> UV irradiance is required, respectively. A pilot scale study conducted by Langmark et al. [93] demonstrated inability of UV irradiation to reduce biofilm associated amoebae. In contrast with other protozoans, members of the *Acanthamoeba* genera are more resistant to both chemical and physical disinfection protocols.

As per water quality guidelines of WHO [65], 41 mg-min/<sup>L</sup> chlorine at 25 ◦C OR 1000 mg-min/<sup>L</sup> monochloramine at 15 ◦C OR 7.3 mg-min/<sup>L</sup> chlorine dioxide 25 ◦C OR 0.63 mg-min/<sup>L</sup> O3 at 15 ◦C OR 10 mJ/cm<sup>2</sup> UV rays, treatments are required for inactivation of pathogenic protozoan (reference protozoa *Giardia*), as mentioned earlier in this section protozoans facilitating growth of *L. pneumophila* can thrive in these conditions (Table 3).

So far, studies have investigated the efficacy of water disinfection protocols against *Acanthamoeba*, *Hartmannella*, *Naegleria* and *Vermamoeba*. However, there are numerous other waterborne cyst-forming, non-cys<sup>t</sup> forming and ciliated protozoans which support the proliferation of *L. pneumophila*. Therefore, there is a need for more research and a standardized approach to evaluating disinfection protocol(s) that target both *L. pneumophila* and potential protozoan hosts. According to our literature survey, the effectiveness of available disinfection protocols depends upon the species, strain and cellular state of protozoans, as well as the type of disinfection technique and exposure time.

## *3.3. Detection Methods*

The most commonly used methods to investigate potential *L. pneumophila* protozoan hosts are co-culture and co-isolation assays [19]. The co-culture assay is widely used in the laboratory to study *Legionella*-protozoan interactions. In this method, *Legionella* is allowed to grow in a protozoan host and fate of bacterium is determined microscopically [94]. In vitro laboratory studies showed that *Acanthamoeba* [95] and *Tetrahymena* [96] allow intracellular replication and packaging of live *L. pneumophila* into export vesicles. Other protozoan genera; *Balamuthia* [97], *Dictyostelium* [98], *Echinamoeba* [31], *Naegleria* [99], *Paramecium* [100], and *Vermamoeba* [32], facilitate intracellular replication of *L. pneumophila*. The second method is used to detect naturally co-existing *Legionella*-protozoans from environment, but microscopically it is very difficult to find protozoans containing *Legionella* in the natural environment [101]. As an alternative approach, a sample is screened for the presence of both *Legionella* and protozoan hosts. Generally, samples are screened by PCR [102,103], fluorescence in situ hybridization [104], classical culturing techniques and microscopy [105,106]. These methods are good for screening environmental samples but are unable to delineate the underlying interactions between *Legionella* and host protozoans. Nowadays, PCR based protocols are widely used to detect *L. pneumophila* and protozoan hosts from engineered water systems. In comparison to classical culturing methods, these protocols are rapid and highly sensitive. However, most of the nucleic acid-based protocols are unable to differentiate viable and dead organisms. Propidium monoazide-PCR or ethidium monoazide-PCR are modified nucleic acid detection protocols to enumerate the live bacteria [107,108] and protozoan hosts [109,110]. To estimate burden of *L. pneumophila* and protozoan hosts in water distribution system, it is necessary to measure the quantity of alive and dead organisms regularly. This literature review demonstrates that *Vermamoeba* and *Acanthamoeba* are predominant hosts of *L. pneumophila* in the context of hospital and potable water systems. Many cyst-forming, non-cys<sup>t</sup> forming and ciliated protozoans have been found associated with *L. pneumophila* and are identified as potential hosts; however, in vitro co-culture assays and microscopic studies are required for confirmation and characterization of this interaction.

During stress (i.e., thermal, nutrient, chemical and radiation), *L. pneumophila* can enter into a VBNC state. After the end of such a stress period, in presence of a suitable host or favorable environmental conditions, the VBNC state can transform back into metabolically active cellular state [111]. Importantly, the underlying mechanisms of resuscitation from VBNC are not ye<sup>t</sup> well understood. However, as the VBNC form is by definition a non-culturable state, classical microbiology culturing techniques cannot be used to monitor viability. Thus, in vitro co-culture assays can be used to resuscitate VBNC in the laboratory [74]. Alternative approaches to analyze VBNC are the analysis of membrane integrity and molecular screening [112]. There are also studies that have shown that intracellular replication of *L. pneumophila* induces VBNC state. According to Buse et al. [26] transformation of *V. vermiformis* trophozoites into cysts promotes biogenesis of VBNC *L. pneumophila*. Therefore, the interaction with protozoan hosts may also affect the ability to monitor the efficacy of disinfection protocols against *L. pneumophila*, because the bacteria may be in the VBNC form. Available literature only discusses disinfection protocols, which target culturable *L. pneumophila*. To our knowledge, there are limited studies investigating the effectiveness of disinfection protocols to eliminate VBNC *L. pneumophila*. It is our suggestion to adopt membrane integrity and in vitro co-culture assays to evaluate the disinfection procedure against VBNC *L. pneumophila*.





1 Most of the studies focus on culturable bacteria, and non-culturable cells are not estimated. 2 No further bacterial inactivation possible, 1–2 × 102 CFU/mL *L. pneumophila* remain stable. 3 Experiments conducted on *Legionella* sp.

#### **4. Materials and Methods**

The databases Scopus and Web of Science were searched for articles written in English containing the keywords ("*Legionella pneumophila*" OR "*L. pneumophila*") AND (*Acanthamoeba* OR *Vermamoeba* OR *Hartmannella* OR *Dictyostelium* OR *Naegleria* OR *Tetrahymena* OR *Echinamoeba* OR *Paramecium* OR *Balamuthia* OR *Oxytricha* OR *Stylonychia* OR *Diphylleia* OR *Stenamoeba* OR *Singhamoeba* OR *Filamoeba* OR Protozoa OR Protozoan OR Amoeba). The above search terms were modified from the review conducted by Boamah et al. [19]. Figure 1 presents the systematic approach to article inclusion or exclusion. Articles were screened by reading the titles and abstracts and initially excluded if they did not refer to a study that detected *L. pneumophila* and a potential protozoan host from a hospital or potable/drinking water source. Articles were then read in full and excluded if they only described laboratory based simulated or pilot-scale experiments on registered bacterial and protozoan strains.
