**An OMV-Based Nanovaccine Confers Safety and Protection against Pathogenic** *Escherichia coli* **via Both Humoral and Predominantly Th1 Immune Responses in Poultry**

**Rujiu Hu 1, Haojing Liu 1, Mimi Wang 1, Jing Li 2, Hua Lin 1, Mingyue Liang 1, Yupeng Gao 1,\* and Mingming Yang 1,\***


Received: 6 September 2020; Accepted: 16 November 2020; Published: 20 November 2020

**Abstract:** Avian pathogenic *Escherichia coli* (APEC) infection in poultry causes enormous economic losses and public health risks. Bacterial outer membrane vesicles (OMVs) and nano-sized proteolipids enriched with various immunogenic molecules have gained extensive interest as novel nanovaccines against bacterial infections. In this study, after the preparation of APEC O2-derived OMVs (APEC\_OMVs) using the ultracentrifugation method and characterization of them using electron microscopy and nanoparticle tracking analyses, we examined the safety and vaccination effect of APEC\_OMVs in broiler chicks and investigated the underlying immunological mechanism of protection. The results showed that APEC\_OMVs had membrane-enclosed structures with an average diameter of 89 nm. Vaccination with 50 μg of APEC\_OMVs had no side effects and efficiently protected chicks against homologous infection. APEC\_OMVs could be effectively taken up by chicken macrophages and activated innate immune responses in macrophages in vitro. APEC\_OMV vaccination significantly improved activities of serum non-specific immune factors, enhanced the specific antibody response and promoted the proliferation of splenic and peripheral blood lymphocytes in response to mitogen. Furthermore, APEC\_OMVs also elicited a predominantly IFN-γ-mediated Th1 response in splenic lymphocytes. Our data revealed the involvement of both non-specific immune responses and specific antibody and cytokine responses in the APEC\_OMV-mediated protection, providing broader knowledge for the development of multivalent APEC\_OMV-based nanovaccine with high safety and efficacy in the future.

**Keywords:** avian pathogenic *Escherichia coli*; nanovaccine; outer membrane vesicles; immune response; broiler

#### **1. Introduction**

Avian pathogenic *Escherichia coli* (APEC) is one of the major pathogens that have been recognized as serious threats to the global poultry industry [1–3]. APEC causes a variety of local and systemic diseases in many avian species, such as colibacillosis in broiler chickens [2,4]. Chicken colibacillosis is characterized by high morbidity and mortality, leading to substantial economic losses every year in the poultry industry worldwide [5]. In commercial production, antibiotic regimens are used as a common measure to control APEC. However, the prevalence of multidrug-resistant APEC caused by the extensive usage of antibiotics has attracted significant concerns. In many countries, antibiotics have

been banned in the animal food industry [6]. Furthermore, drug residues and transfers of resistant genes through poultry products are becoming severe threats to public health [7]. Therefore, it is necessary to explore novel preventive approaches. To date, the use of effective vaccines is recognized as an important way to control APEC infections in today's large-scale poultry industry [8,9].

Many attempts have been made to develop various biological materials as vaccine candidates against APEC infections. Some cell-wall components and virulence factors from APEC strains, such as lipopolysaccharide (LPS), outer membrane proteins, siderophore receptor protein, fimbriae and adhesins, have been shown to induce protective immunity against their corresponding serotypes [8,10,11]. Accordingly, numerous APEC vaccines, mainly inactivated, live-attenuated and subunit vaccines, have been developed for commercial use. Although these vaccines have been proven to be effective, there are still some drawbacks in practical application [8]. Inactivated vaccines are prepared by inactivating the live whole-bacteria with heat or chemicals, which can provide short-term protection against the homologous serogroups only [8]; live-attenuated bacteria vaccines may cause public safety concerns due to their potential risk of bacterial spread [9]. Subunit or recombinant vaccines, mainly including iron regulated outer membrane proteins-based vaccines, fimbriae-based vaccines and increased serum survival protein-based vaccines, could induce better protective immunity against heterologous challenges than inactivated vaccines, but they are rarely used in practice because of their limitations, such as unstable efficacy and the requirement of strong adjuvants [8,9]. Additionally, APEC isolates commonly have a variety of O serogroups (according to the O-antigens); and three main serogroups, including O1, O2 and O78, are frequently associated with disease formation in poultry farms, which can cause over 80% of chicken colibacillosis cases [2,4,8]. It may be difficult to achieve better prevention efficiency for APEC multi-serogroups using these above-mentioned vaccines. Therefore, it is still necessary to develop new vaccine candidates with both higher safety and better efficacy.

Almost all domains of life, including bacteria, archaea and eukaryotes, can secrete nanosized membrane vesicles during their normal growth [12]. These nanovesicles released by Gram-negative bacteria originate from the outer membrane of the cell envelope, and thus are also termed outer membrane vesicles (OMVs) [13]. Biochemical and proteomic analyses have shown that OMVs are naturally enriched with many bioactive molecules of the parental bacteria, including outer membrane proteins and lipids, periplasmic proteins, polysaccharides, nucleic acids (DNA and RNA) and virulence-associated factors [14–17]. Bacterial OMVs have been proven to play important roles in host–bacteria interactions, such as mediating pathogenesis, enhancing bacterial survival under various environmental stress conditions and modulating host immunity [18,19]. Due to the unique structural and immunological properties, such as biocompatible nanometer-scale structure, and the feature of being genetically modified and naturally carrying both adjuvants and multiple antigens, OMVs are generally considered to be emerging candidates for drug delivery platforms and nanovaccines [20–23]. Numerous studies have demonstrated that OMVs secreted by a variety of pathogens, such as *Neisseria meningitidis* [24], *Klebsiella pneumoniae* [25], *Vibrio cholerae* [26] *Bordetella pertussis* [27], *Salmonella* [28] and *Staphylococcus aureus* [29], can elicit protection against the corresponding bacterial infections in mice. Moreover, some studies have shown that immunization with OMVs can provide broad cross-protection against heterologous serogroups [30–32].

Although extensive studies have revealed that OMVs secreted by pathogenic *E. coli* species can induce protective immunity in mouse models, very few investigations have focused on whether OMVs produced by APEC (APEC\_OMVs) have the potential to be developed as a novel vaccine candidates in chickens [21,33,34]. Recently, Wang and colleagues have revealed that vaccination with APEC O78-derived OMVs can protect broiler chickens against homologous infection, suggesting the potential of APEC-derived OMVs as APEC vaccine candidates [35]. Wang's report only characterized the specific antibody responses induced by OMVs; however, the safety of APEC\_OMVs and the protective mechanisms involving both innate and specific cellular immunity have not been clearly identified.

In this study, we isolated and purified APEC\_OMVs from a clinical APEC O2 strain, a major APEC serogroup causing chicken colibacillosis. Compared with Wang's study, the present study investigated the immunogenicity of APEC\_OMVs in a broiler chick model using both in vitro and in vivo experiments, including innate immune responses in chicken macrophages; non-specific immune factor activities and specific antibody responses in the serum; and lymphocyte proliferation and cytokine responses in splenic lymphocytes. Moreover, we also evaluated the adverse effects of APEC\_OMVs and estimated the window dose between effectiveness and toxicity for APEC\_OMVs. Our work reveals the detailed immunologic mechanisms of APEC\_OMV-mediated protection, providing the basic information for the development of an effective and multivalent APEC\_OMV-based nanovaccine in the future.

#### **2. Materials and Methods**

#### *2.1. Bacterial Strain and Preparation of APEC\_OMVs*

A clinical APEC O2 strain from a chicken with colisepticemia (collection number CVCC1554) was purchased from China Veterinary Culture Collection Center (China Veterinary Drug Supervision Institute, Beijing, China) and used in this study. This APEC isolate was grown in Luria–Bertani (LB) broth at 37 ◦C. Native OMVs were isolated and purified from bacterial culture supernatant by a series of centrifugal processes, as described in our previous studies [36,37]. Briefly, the bacteria-free supernatant was collected from the culture medium in the logarithmic phase by centrifugation (12,000× *g*, 15 min, 4 ◦C), and filtered through a 0.45-μm membrane (Merck Millipore, Tullagreen, Carrigtwohill, Ireland) followed by ultracentrifugation (150,000× *g*, 2 h, 4 ◦C). After washing with sterile phosphate buffer saline (PBS; pH 7.4), the obtained APEC\_OMV pellet was purified by discontinuous density centrifugation. For purification, APEC\_OMVs were covered with 20% (1.127 g/mL) and 35% (1.199 g/mL) OptiPrep (Sigma, catalogue number D1556) and then subjected to ultracentrifugation (16 h, 180,000× *g*, 4 ◦C) [38]. The interlayer of the 20% and 35% OptiPrep containing the majority of vesicles was collected, dispersed in sterile PBS and then centrifuged (150,000× *g*, 2 h, 4 ◦C) to completely remove OptiPrep. The purified APEC\_OMVs were uniformly dispersed in sterile PBS and any bacterial contaminations were removed by filter sterilization with a 0.45-μm membrane. The APEC\_OMVs samples were stored at −80 ◦C for future use. The protein quantification of APEC\_OMVs was performed using a TaKaRa BCA Protein Assay Kit (TaKaRa Bio, Beijing, China; catalogue number T9300A) following the manufacturer's instructions.

#### *2.2. Electron Microscopy Analysis*

APEC\_OMVs were visualized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). For SEM visualization, 10 μL of the purified APEC\_OMVs was dropped on a 5 mm × 5 mm silicon slice, dried at 20 ◦C and sputter-coated with gold-palladium using an ion-sputtering coater (E-1045; Hitachi, Tokyo, Japan). The prepared nanovesicles were observed using a Field Emission Scanning Electron Microscope (S-4800, Hitachi, Tokyo, Japan). For TEM visualization, 200 μg/mL of the nanovesicles was adhered to 300-mesh copper grids for 10 min followed by negatively staining with 1% phosphotungstic acid (pH 7.2), and then viewed using FEI Tecnai™ G2 Spirit BioTWIN (FEI Company, Hillsboro, OR, USA) at 100 kV.

#### *2.3. Nanoparticle Tracking Analysis*

Nanoparticle tracking analysis (NTA) was performed to measure the diameter size and particle number of APEC\_OMVs using an NS300 nanoparticle analyzer (Malvern, Worchestershire, UK). These nanovesicle samples were uniformly dispersed in PBS and detected with a camera level of 15. Five 60 s video records were obtained for each sample and analyzed using NTA software version 2.3. The detection threshold was set at 6.

#### *2.4. Animals and Housing*

Broiler chicks (Arbor Acres) were hatched from fertilized eggs in an automatic incubator (Beijing LanTianJiao Electronic Technology Co., Ltd., Beijing, China) following routine incubation procedures in a sterilized room with filtered air. All hatching eggs and the incubator were sterilized before incubation. The hatching eggs were sterilized by wiping the surface of the eggshell with 75% alcohol cotton balls before putting them into the incubator. The incubator was placed in an isolation room and the isolation room was disinfected by formaldehyde fumigation. Newly hatched chicks were housed in stainless-steel cages in sterilized rooms with filtered air, strict sanitary conditions and age-appropriate temperatures. All chicks were fed an age-appropriate commercial diet containing no antibiotic additives. Drinking water and diets were offered ad libitum. All procedures of animal experiments were approved by the Ethics Committee of Animal Care and Use at Northwest A&F University with the permit number 2018NWAFU-052.

#### *2.5. Maternal Anti-APEC Antibody Levels in Broiler Chicks*

The objective of this experiment was to detect the optimal age for the immunization in young broiler chicks. A total of 30 newly hatched chicks were randomly divided into 6 replicates with 5 birds per replicate. The serum samples were collected at 1, 3, 5, 7, 10, 14, 18 and 21 days of age to determine the natural anti-APEC maternal antibody levels as described below.

#### *2.6. E*ff*ect of APEC\_OMV Vaccination on the Growth Performance, Immune Organ Index and Blood Cell Counts*

This experiment aimed to evaluate the potential adverse effects of APEC\_OMVs. For vaccination procedures, a total of 120 seven-day-old broiler chicks were randomly divided into four groups. Each group contained 6 replicates with 5 birds per replicate, which were respectively immunized with 200 μL PBS (as a control) and 10, 50 and 200 μg of APEC\_OMVs in 200 μL PBS via intramuscular injection into the right thigh muscle using the disposable syringe at 7 and 14 days of age. The body weight, feed intake and the number of deaths were recorded on a replicate basis and used to calculate average daily weight gain (ADWG), average daily feed intake (ADFI) and feed conversion rate (FCR) from 7 to 21 days of age. At 21 days of age (one week after the secondary vaccination), 6 chicks of each group were chosen and euthanized to collect immune organs (thymus, spleen and bursa of Fabricius) and blood samples for the determination of immune organ index and blood cell counts. The organ index was calculated based on the following formula: organ index = organ weight (g)/body weight (kg). The numbers of red blood cells (RBC) and white blood cells (WBC) were estimated by a manual hemocytometer using Natt–Herrick's stain solution [39].

#### *2.7. Effect of APEC\_OMV Vaccination on the Protective Efficacy against Homologous Infection in Broiler Chicks*

After the vaccination procedures, the remaining 24 chicks of each group were challenged by the air sac route with 5 <sup>×</sup> <sup>10</sup><sup>8</sup> CFU/bird of APEC O2 recommended by the previous study at 21 days of age [10]. The survival rate of chicks in each group was calculated daily for 10 consecutive days. Blood samples were collected from PBS- and APEC\_OMV-immunized chicks at 12, 24 and 36 h after bacterial challenge for the determination of bacterial loads. Blood samples were prepared by 10-fold serial dilution in sterile PBS, followed by plating on LB agar plates in triplicate. The counts of bacterial colonies under 37 ◦C for 12 h were recorded. Serum samples were collected from PBS- and APEC\_OMV-immunized chicks at 24 h after bacterial challenge for the determination of proinflammatory cytokines interleukin (IL)-1β and IL-6 using the Chicken Interleukin 1β ELISA Kit (Cloud-Clone Corp., Houston, TX, USA; catalogue number SEA563Ga) and Chicken Interleukin 6 ELISA Kit (Cloud-Clone Corp., Houston, TX, USA; catalogue number SEA079Ga) according to the manufacturer's instructions, respectively.

#### *2.8. In Vitro Chicken Macrophage Assays*

HD11 cells, a chicken macrophage cell line derived from bone marrow [40], were used in this study and cultured in the complete PRMI-1640 medium (Gibco, catalogue number 22400089) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Zeta-Life, catalogue number Z7181FBS-500), 100 U/mL penicillin and 100 μg/mL streptomycin (Sigma, catalogue number P4333) in an atmosphere of 5% CO2 at 37 ◦C. APEC\_OMVs (5 μg/mL) were stained with 1 μM dialkylcarbocyanine iodide (DiI; Sigma, catalogue number 42364) as described previously [41]. The DiI-labeled APEC\_OMVs were co-incubated with HD11 cells (5 <sup>×</sup> <sup>10</sup><sup>6</sup> cells/well) in a 24-well culture plate. After 4 h-incubation, the cells were collected, washed and then fixed with 4% paraformaldehyde in PBS followed by cell nucleus staining with 10 μg/mL of 4 ,6-diamidino-2-phenylindole (DAPI; Sigma, catalogue number D9542). Subsequently, the samples were placed on glass slides and viewed by an Andor Revolution XD spinning-disk confocal microscope (Andor Technology, UK).

To investigate innate immune responses induced by the APEC\_OMVs, HD11 cells (5 <sup>×</sup> <sup>10</sup><sup>6</sup> cells/well) were stimulated with serial doses of APEC\_OMVs (0–1000 ng/mL) in a 24-well culture plate. After a 16 h treatment, the cells were harvested for the measurement of the expression of major histocompatibility complex class II β (MHC IIβ) and cytokines tumor necrosis factor α (TNF-α) and IL-6 by quantitative real-time PCR (qRT-PCR).

#### *2.9. Serum Non-Specific Immune Factor Activities*

Serum samples were collected from PBS- and APEC\_OMV-immunized chicks at 14 and 21 days of age (one week after the primary and secondary vaccinations, respectively) before APEC\_OMV vaccination or bacterial challenge. The activities of lysozyme and superoxide dismutase (SOD) in the serum were measured using commercial assay kits (Nanjing Jiancheng Bioengineering Institute, Jiangsu, China; catalogue number A050-1-1 and A001-3-2). Serum complement 3 level was estimated using a chicken-specific ELISA kit (Cloud-Clone Corp., Houston, TX, USA, catalogue number SEA861Ga). The respiratory burst activity in blood leukocyte was estimated as described previously [42].

#### *2.10. Determination of Specific Antibody Titer and Bactericidal Activity in Serum*

The natural antibody levels in nonimmunized serum and the levels of APEC\_OMV-reactive IgY and APEC-reactive IgY in PBS- and APEC\_OMV-immunized serums were determined by the ELISA method as previously described [35]. Bacterial lysates or APEC\_OMVs (200 ng/well) were used as antigens and the diluted serum samples (1:200) were used as the primary antibodies. After incubation with antigens, the primary antibodies were reacted with the secondary horseradish peroxidase-conjugated rabbit anti-chicken IgY (200 ng/mL; abcam, catalogue number ab97140) followed by termination with substrate tetramethylbenzidine (100 μL). The OD450 was detected using a BioTek synergy2 microplate reader (Biotek, Winooski, VT, USA).

To further evaluate serum antibody responses induced by APEC\_OMVs, the ability of bacteria-killing by PBS- and APEC\_OMV-immunized serum samples was estimated according to the previously described method [35]. The bacterial survival rate was calculated by comparing bacterial colony-forming unit (CFU) counts after and before the treatment with the serum. Each sample was detected in triplicate.

#### *2.11. Lymphocyte Proliferation Assays*

Splenic and peripheral blood lymphocytes were freshly isolated from PBS- and APEC\_OMVimmunized chicks at 21 days of age (one week after the secondary vaccination) using the Chicken Splenic and Peripheral Blood Lymphocyte Isolation Kits (Solarbio, catalogue number P9120 and P8740), respectively, following the manufacturer's instructions. These isolated lymphocytes were maintained in complete PRMI-1640 medium supplemented with 10% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin in an atmosphere of 5% CO2 at 37 ◦C. Cell proliferation responses were detected by

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay using an MTT Cell Growth Assay Kit (Sigma, catalogue number CT02) as previously described [43]. The results of the tests were expressed as proliferation index according to the formula: proliferation index = (OD570 of test sample-OD570 of the negative control)/OD570 of the negative control. Each sample was determined in triplicate.

#### *2.12. Re-Stimulation Assay of Splenic Lymphocyte*

Splenic lymphocytes were obtained and cultured as described above. These cells were adjusted to a concentration of 5 <sup>×</sup> 106 cells/well in a 24-well culture plate and then re-stimulated with APEC\_OMVs (5 μg/mL) for 12 h. The cells were harvested to measure the gene expressions of IFN-γ, IL-4 and IL-17A.

#### *2.13. Quantitative Real-Time PCR (qRT-PCR) for mRNA Quantification*

Total RNA was extracted from the cultured HD11 cells and splenic lymphocytes using a Total RNA Kit (Omega Bio-Tek, catalogue number R1034) following the manufacturer's instructions. After examination of RNA purity and quality with an ND-1000 spectrophotometer (Nano-drop Technologies, Wilmington, Delaware), the qRT-PCR analysis was performed with One Step TB Green® PrimeScript™ PLUS RT-PCR Kit (TaKaRa Bio, Beijing, China; catalogue number RR096A) on an iCycler IQ5™ Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) according to the manufacturer's instructions. The primer sequences of the target genes and a housekeeping gene (β-actin) are shown in Supplementary Table S1. Triplicate qRT-PCR reactions for each sample were conducted with the following protocol: 95 ◦C for 1 min; 40 cycles of 95 ◦C for 15 s and 60 ◦C for 30 s, 72 ◦C for 30 s; 72 ◦C for 10 min. Relative mRNA expression was calculated using the 2−ΔΔCt method as described previously [44], and expressed as the fold-change relative to the control, which was normalized to 1.

#### *2.14. Statistical Analysis*

Experimental data were shown as mean ± standard error (S.E.). Data were analyzed by Graph Pad Prism software 5.0 (San Diego, CA, USA). The analysis of differences between the two groups was performed by Student's *t*-test. The analysis of differences among greater than two groups was performed by one-way ANOVA analysis with the Newman–Keuls post-test. The survival data after bacterial infection were analyzed by the log-rank test. Statistical significance was declared at *P* < 0.05.

#### **3. Results**

#### *3.1. Characterization of APEC\_OMVs*

APEC\_OMVs were obtained from the culture supernatant of APEC O2 strain using a series of centrifugation procedures and then purified using the density gradient centrifugation method. The morphology and integrity of APEC\_OMVs were detected by SEM (Figure 1A) and TEM (Figure 1B). The results showed that these vesicles were intact nanosized structures with spherical morphology, and the majority of these nanovesicles ranged from 50 to 150 nm in diameter. Typical results characterized by NTA are shown in Figure 1C,D. The diameter of APEC\_OMVs peaked at 89 nm, which is in accordance with the results of electron microscopy analyses and the previously determined sizes of bacterial OMVs [34].

**Figure 1.** Characterization of outer membrane vesicles secreted by avian pathogenic *E. coli* O2 (APEC\_OMVs). (**A**) The purified APEC\_OMVs were viewed with a scanning electron microscope. (**B**) After negative staining with 1% phosphotungstic acid, APEC\_OMVs were visualized by transmission electron microscopy. The red arrows indicate several visible APEC\_OMVs. (**C**) The image from the movie captured using a SCMOS camera of Malven NTA 3.0 when APEC\_OMVs were characterized by nanoparticle tracking analysis (NTA). (**D**) Concentration and size distribution of APEC\_OMVs determined by NTA.

#### *3.2. Natural Antibody Levels in Nonimmunized Chicks*

As shown in Figure 2, a high antibody level was observed in the serum before 5 days of age. The antibody titer dropped to a low level at 7 days of age and remained stable thereafter. These findings suggested that the optimal age of vaccination should not be earlier than 7 days of age.

**Figure 2.** Natural anti-APEC maternal antibody levels in the serum collected from commercial broiler chicks during the 1–28 days of age (n = 5). Data are presented as mean ± SE.

#### *3.3. Effect of APEC\_OMVs Vaccination on the Growth Performance, Immune Organ Index and Blood Cell Counts*

As shown in Table 1, vaccination with 10 and 50 μg of APEC\_OMVs had no significant impacts on ADFI, ADWG, FCR, immune organ index or WBC and RBC counts. However, vaccination with 200 μg of APEC\_OMVs significantly reduced ADFI and ADWG, and increased FCR and WBC count, suggesting that this vaccination dose can have adverse effects on chicks. The immune organ indexes were slightly improved in all APEC\_OMV-immunized groups compared with the control group, but an increase in the dose of APEC\_OMVs to 200 μg had no statistically significant effect on the immune organ indexes. No chicks died in all groups during the vaccination period (data not shown). These results demonstrated that vaccination with an appropriate dose of APEC\_OMVs was safe, while high doses of APEC\_OMVs could be toxic.


**Table 1.** Effects of various doses of APEC\_OMV vaccination on the growth performance, immune organ index and blood cell counts.

<sup>1</sup> APEC\_OMVs = avian pathogenic *Escherichia coli* O2-derived outer membrane vesicles. <sup>2</sup> ADFI = average daily feed intake; ADWG = average daily weight gain; FCR = feed conversion rate; WBC = white blood cells; RBC = red blood cells. <sup>3</sup> SE = Standard error of the mean. Growth performance were calculated from 7 to 21 days of age; immune organ index and blood cell counts were measured at 21 days (one week after the secondary vaccination). a,b Different superscript letters in the same row indicate significant difference (*p* < 0.05).

#### *3.4. Vaccination with APEC\_OMVs Was Protective against Homologous Infection in Broiler Chicks*

The procedures for the vaccination and bacterial challenge are shown in Figure 3. As shown in Figure 4A, only 16.7% of chicks in the PBS-immunized group survived 10 days after the bacterial infection. The survival rates of 45.8% and 83.3% were observed for groups immunized with 10 and 50 μg of APEC\_OMVs, respectively, which were significantly higher than those in the PBS-immunized group. However, an increase of the vaccination dose to 200 μg did not result in significant improvement of protective efficacy. Together with the results of experiment 2, the dose of 50 μg was selected as the final dosage for the following analyses. To further confirm the protective efficacy conferred by APEC\_OMVs, bacterial loads in blood samples and proinflammatory cytokine production in serum samples were determined after bacterial challenge. As illustrated in Figure 4B, the APEC\_OMV-immunized group showed significantly lower bacterial counts in the blood after 24 h post-infection compared with the control group, indicating that effective clearance of bacteria was induced by APEC\_OMVs. Additionally, the levels of proinflammatory cytokines IL-1β and IL-6 in the APEC\_OMV-immunized serum were significantly lower than those in the PBS-immunized serum (Figure 4C). These results revealed that vaccination with APEC\_OMVs could reduce bacterial loads and proinflammatory cytokine production, and thus provided protection against homologous challenge.

**Figure 3.** Timeline of APEC\_OMV vaccination and bacterial challenge in broiler chick experiments. Four groups of broiler chicks (n = 30; 6 replicates per group with 5 birds per replicate) were intramuscularly immunized with PBS (as a control) and various doses of APEC\_OMVs at 7 and 14 days of age, respectively, and challenged by the air sac route at 21 days of age. Blood and spleen samples were collected from PBS- and APEC\_OMV-immunized chicks at indicated times.

**Figure 4.** Protective efficacy conferred by APEC\_OMV vaccination against homologous challenge in broiler chicks. (**A**) Survival rates of PBS- and APEC\_OMV-immunized birds after bacterial challenge (n = 24). The number of surviving chicks in each group was recorded daily for 10 consecutive days after bacterial challenge. (**B**) Bacterial loads in peripheral blood collected from PBS- and APEC\_OMV (50 μg)-immunized chicks at indicated times after bacterial challenge (n = 6). (**C**) The production of proinflammatory cytokines (IL-1β and IL-6) in the serum collected from PBS- and APEC\_OMV (50 μg)-immunized chicks at 22 days of age (24 h after bacterial challenge) (n = 6). Data are presented as mean ± SE. \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001, versus the respective control.

#### *3.5. APEC\_OMVs Activated Innate Immune Responses In Vitro*

Macrophages are not only important players in the clearance of pathogens but also function as antigen-presenting cells (APCs) to recognize and process foreign antigens, linking the innate and adaptive immune responses. We first examined whether chicken HD11 macrophages recognize and respond to APEC\_OMVs in vitro. After co-incubation with HD11 cells for 4 h, these DiI-labeled vesicles (red signals) were observed in the cytoplasms of these cells, suggesting that APEC\_OMVs were internalized by HD11 cells (Figure 5A). Furthermore, stimulation with APEC\_OMVs dramatically enhanced the expression of MHC IIβ and cytokines TNF-α and IL-6 in a dose-dependent manner (Figure 5B). These results indicated that APEC\_OMVs could provoke innate immune responses in APCs, which can activate T cell responses.

**Figure 5.** Innate immune responses induced by APEC\_OMVs in chicken macrophages in vitro. (**A**) Visualization of internalization of APEC\_OMVs by chicken HD11 macrophages using a confocal microscope. Row 1 was a non-stimulated treatment. Row 2 was stimulated with 5 μg/mL of DiI-labeled APEC\_OMVs (red signal) for 6 h at 37 ◦C followed by cell nucleus staining with 10 μg/mL of DAPI (blue signal). (**B**) qRT-PCR analysis for the expression of MHC IIβ, TNF-α and IL-6 in HD11 cells stimulated with various concentrations of APEC\_OMVs for 16 h. Results are representatives of three independent experiments and expressed as mean ± SE. \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001, versus the control.

#### *3.6. Vaccination with APEC\_OMVs Improved Serum Non-Specific Immune Factor Activities*

We next evaluated the vaccination effect of APEC\_OMVs on the serum non-specific immune factor activities. As shown in Figure 6, vaccination with 50 μg of APEC\_OMVs significantly improved the lysozyme, complement 3, respiratory burst and SOD activities after both primary and secondary vaccinations. Moreover, the activity of these non-specific immune factors was significantly higher after the secondary booster vaccination than after the primary vaccination.

**Figure 6.** Effects of APEC\_OMV vaccination on non-specific immune factor activities in the serum. Serum samples were collected from PBS- and APEC\_OMV (50 μg)-immunized chicks at 14 and 21 days of age (one week after the primary and secondary immunization, respectively) before the secondary APEC\_OMV immunization or bacterial challenge to determine the activities of the following immune factors: (**A**) lysozyme; (**B**) complement 3; (**C**) respiratory burst; (**D**) SOD. Data are presented as mean ± SE (n = 6). \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001.

#### *3.7. APEC\_OMV-Induced Protection Was Associated with Elevated Antibody Responses*

To identify adaptive immune responses involved in APEC\_OMV-induced protection, we first detected the specific antibody responses one week after the primary and secondary vaccinations. The levels of APEC\_OMV-reactive IgY and APEC-reactive IgY were significantly elevated when chicks were immunized with APEC\_OMVs (Figure 7A,B). The secondary booster vaccination significantly enhanced the antibody responses compared to the primary vaccination. Furthermore, APEC\_OMV-immunized serum samples showed higher bactericidal activities after both primary and secondary vaccinations compared with PBS-immunized serum samples (Figure 7C).

**Figure 7.** Specific IgY levels and bactericidal activities of serum samples from APEC\_OMV-immunized chicks. Serum samples were collected from PBS- and APEC\_OMV (50 μg)-immunized chicks at 14 and 21 days of age (one week after primary and secondary vaccination, respectively) before the secondary APEC\_OMV vaccination or bacterial challenge. The production of APEC\_OMV-reactive IgY (**A**) and APEC-reactive IgY (**B**) was estimated by ELISA assays. (**C**) Bacteria killing assay by the APEC\_OMV-immunized serum. The APEC O2 strain was incubated with PBS- and APEC\_OMV (50 μg)-immunized serum samples at 37 ◦C for 1 h and then the bacterial survival rate was measured. Data are presented as mean ± SE (n = 6). \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001.

#### *3.8. Vaccination with APEC\_OMVs Induced Lymphocyte Proliferation and a Predominant Th1 Response*

We finally identified cellular responses associated with protection conferred by APEC\_OMVs. As shown in Figure 8A, APEC\_OMV vaccination significantly enhanced the proliferation of both spleen lymphocytes and peripheral blood lymphocytes in response to mitogen. To determine which T-cell responses were involved in APEC\_OMV-induced protection, spleen lymphocytes were isolated from the immunized chicks one week after the secondary vaccination and re-stimulated with APEC\_OMVs. The results showed that re-stimulation with APEC\_OMVs significantly upregulated the expression of IFN-γ (a representative Th1 cytokine) and IL-17A (a representative Th17 cytokine); the expression of IL-4 (a representative Th2 cytokine) remained unchanged between PBS- and APEC\_OMV-immunized groups. Meanwhile, the degree of IFN-γ upregulation (over 12-fold change compared with the control) was much higher than that of IL-17A upregulation (only 3.8-fold change compared with the control), suggesting that IFN-γ-mediated Th1 response may play a predominant role.

**Figure 8.** Lymphocyte proliferation and cytokine responses induced by APEC\_OMVs. At 21 days of age (one week after secondary vaccination) before the bacterial challenge, splenic lymphocytes and peripheral blood lymphocytes were isolated from PBS- and APEC\_OMV (50 μg)-immunized chicks. (**A**) Proliferation indexes of MTT assays for the splenic lymphocytes and peripheral blood lymphocytes re-stimulated with mitogen (ConA). (**B**) Analysis of mRNA expression of Th1 cytokine (IFN-γ), Th2 cytokine (IL-4) and Th 17 cytokine (IL-17) in the splenic lymphocytes re-stimulated with APEC\_OMVs (5 μg/mL) for 24 h. Data are presented as mean ± SE. \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001; NS, not significant.

#### **4. Discussion**

Vaccination has proved to be the most practical and effective strategy to control bacterial infections. Bacterial OMVs are becoming increasingly attractive as effective immune-stimulating materials for the development of novel vaccines and adjuvants [20]. Avian colibacillosis, caused by APEC strains, is one of the most severe diseases leading to large economic losses in the global poultry industry [45]. In the present study, we confirmed that vaccination with 50 μg of APEC\_OMVs showed no adverse effects and effectively protected chicks against homologous APEC infection by reducing bacterial loads and proinflammatory cytokine production. These protective effects were mediated through activations of both innate and adaptive immune responses elicited by APEC\_OMVs, which mainly included serum non-specific immune factors, a specific antibody-mediated humoral immune response and an IFN-γ-mediated cellular immune response.

OMV-based vaccines hold several advantages over the current inactivated and attenuated live APEC vaccines. First, a large number of studies and reviews have revealed that OMVs are membranous vesicles with nanoscale sizes in the range of 20–250 nm, which enables them highly biocompatible and capable of delivering interior molecules in concentrated and protected forms [13,46,47]. Consistent with these previous results, electron microscopic and NTA analyses in our work showed that APEC\_OMVs were intact spherical bilayer nanoparticles with a mean size of 89 nm in diameter. These structural characteristics allow OMVs to be more effectively delivered throughout intracellular compartments and efficiently taken-up by APCs. Second, OMVs carry various immune-stimulating molecules from the outer membrane, such as LPS and outer membrane proteins and lipids [13]. Many of these components are immunogenic and act as natural adjuvants, which have the potential to be both multi-antigen carriers and vaccine adjuvants [33,34,48]. These immunological characteristics of these OMV-based vaccines provide great advantages in vaccine efficacy compared to the current single-antigen vaccines. Third, in addition to being easily prepared from natural bacteria, OMVs can be conveniently loaded with specific antigens using bioengineered bacteria [49,50]. Finally, they showed better safety within a certain dose range than attenuated live vaccines due to their nonliving and acellular features, largely reducing the risk of bacterial transmission. However, it should be noted that injection with high doses of OMVs might be toxic because they contain various virulence-associated factors. In this study, we determined that a dose of 50 μg for APEC\_OMVs not only showed no side effects but also induced effective protection. Therefore, extra considerations should be made in the window between efficacy and toxicity for OMV immunization on animals and humans in the future.

Innate immunity is recognized to be the first line of host defense against pathogen-associated invasion. Previous in vitro studies have demonstrated that OMVs could activate the innate immune responses of dendritic cells and macrophages and induce the production of cytokines that regulate the adaptive immune responses [25,33]. As an essential cellular component in the innate immune system, macrophages are not only directly involved in the elimination of pathogens, but also can be used as APCs to recognize and process antigens [51]. Consistent with these findings, in this study, we observed that APEC\_OMVs were effectively taken-up by the chicken HD11 macrophage cell line in vitro and enhanced the expression of MHC IIβ, TNF-α and IL-6. MHC IIβ, mainly expressed in APCs, can assist these cells in presenting exogenous antigen to CD4 T cells, which can activate specific B- and T-cell immune responses [52]. IL-6 is known to induce adaptive immune responses in mammalian Th17 and play an important role in fighting bacterial invasion [33]. It is important to note, however, that these results were derived from a macrophage cell line, and future studies will be needed to investigate whether similar results are expected from primary cells. Additionally, non-specific immune factors in serum are essential for the elimination of pathogens. Lysozyme plays an essential role in the host's innate immune system against bacterial challenge by cleaving cell wall peptidoglycan [53]. The complement molecule is an important innate immune component that can initiate innate responses and modulate adaptive immune responses [54]. Respiratory burst is an oxygen-dependent mechanism by which neutrophils kill invading pathogens [55]. SOD can effectively alleviate inflammatory responses by reducing oxidative stress during bacterial infection [56]. The activities of these non-specific immune

factors were significantly enhanced during APEC\_OMV vaccination, which may be due to the enhanced pathogen clearance accompanied by the involvement of the increased leukocytes [57].

The adaptive immune response plays a very important role in host defense against bacterial infections. Specific antibody and T-cell immune responses induced by bacterial OMVs have been demonstrated in both in vitro and in vivo mouse studies [25,33,58]. However, it is unclear which immune response plays a major role. Some studies have shown that antibody-mediated humoral immunity is the most important factor [26], while other studies have suggested that cellular immunity mediated by cytokines, especially IFN-γ and IL-17, is essential for the protection [25]. In the current study, the ELISA results showed that vaccination with APEC\_OMVs could elevate the production of anti-APEC\_OMVs and anti-APEC antibodies in the APEC\_OMV-immunized serum. These enhanced responses of specific antibody induced by APEC\_OMVs were also confirmed by bactericidal activity assays. These findings are consistent with the recent study showing that OMVs derived from APEC O78 induced similar protective efficacy in an antibody-dependent manner [35]. However, cytokine-mediated cell responses to APEC\_OMV vaccination have not previously been determined. Previous studies performed in mice have indicated that the IFN-γ-mediated Th1 response plays an important role in the protection against bacterial infections by enhancing the bactericidal activity of phagocytes [33,59]. Moreover, the protective effect of vaccines against bacterial challenge requires Th17-mediated immunity by promoting neutrophil recruitment to the site of infection [60]. Meanwhile, our study indicated that APEC\_OMVs promoted splenic and peripheral blood lymphocyte proliferation during vaccination. We further identified that APEC\_OMVs activated the expression of IFN-γ and IL-17A but not IL-4 in splenic lymphocytes. These results are in accordance with the previous study performed in mice with pathogenic *E. coli*-derived OMVs [33]. It is worth noting that the upregulation of IFN-γ induced by APEC\_OMVs was visibly higher than that of IL-17, implying that the IFN-γ-mediated Th1 response may play a dominant role. However, at present, we could not rule out whether other cell types, such as NK cells and γ-δ T cells, are involved in the protection because these cells can also produce IFN-γ and IL-17. The exact cell responses to APEC\_OMVs require further exploration. Taken together, although further study is needed, our present work reveals that vaccination with APEC\_OMVs induces protective immunity against homologous bacterial infection mainly through the induction of specific antibody and IFN-γ-mediated immune responses.

APEC\_OMV-induced protection was well-evidenced in this study, but it is challenging to identify which specific molecules are most essential because APEC\_OMVs contain a variety of immunogenic components. Many reviews have illustrated that protein is the most important component of bacterial OMVs and mediates multiple functions [13,61]. Large-scale proteomic studies of various Gram-positive bacteria-derived OMVs have indicated that outer membrane proteins account for the majority of vesicular proteins [14,62]. Many outer membrane proteins have been proved to contribute to bacterial pathogenesis and are used as protective antigens to induce effective protection against bacterial infections [63,64]. Several outer membrane proteins with high abundance, mainly OmpA, OmpX and OmpW, are commonly found in OMVs secreted by APEC strains and other *E. coli* and can elicit strong protective immunity [35,65,66]. These conserved outer membrane proteins give APEC\_OMVs the ability to confer a certain degree of cross-protection. The cross-protective effects of several pathogens-derived OMVs have been demonstrated in mice [30,32]. Further investigations are required to confirm whether APEC\_OMVs protect against heterologous infections. Additionally, OMVs carry various non-protein antigens, such as LPS, which also participate in the APEC\_OMV-induced protective immunity [61]. It is the combination of these diverse antigens that give OMVs their extensive immunogenicity.

Although our present study may be useful to developing a new APEC vaccine, there remain several shortcomings which shall lead us to future work. First, APEC\_OMV vaccination using the intramuscular route is less likely to be feasible for practical use in large flocks. Further studies can focus on the development of oral OMV-based nanovaccines for poultry, which may be more competitive and attractive for commercial use in broilers. Second, vaccinating twice for broiler chickens can be costly compared to the currently available vaccine. Third, we did not evaluate the stability and uniformity of APEC\_OMVs preparation. Fourth, we did not investigate whether APEC\_OMVs could provide broad protection against heterologous challenges. It is believed that single-serogroup APEC\_OMVs may not provide enough protective efficacy against chicken colibacillosis caused by multiple APEC serogroups. Therefore, it is still necessary to develop multi-serogroup APEC\_OMVs or bioengineered APEC\_OMVs with highly expressed heterologous antigens as broadly-protective vaccine candidates against multiple APEC serogroups in the future. Finally, some important investigations of APEC\_OMVs as a vaccine candidate were not carried out. For example, antibody responses were not determined after the APEC challenge, the survival of chickens and bacterial load until slaughter age (day 42) were not monitored and the APEC lesions in the internal organs were also not investigated. Further studies, including investigations of the protective efficiency and long-term protection of APEC\_OMVs, more detailed protection mechanisms associated with bacterial clearance in multiple organs, pathological changes of internal organs and cytokine responses, etc., will be considered in our future research programs. In conclusion, we revealed that vaccination with APEC\_OMVs protected broiler chicks against homologous infection by enhancing bacterial clearance and reducing proinflammatory cytokine production. We further demonstrated that APEC\_OMVs could activate innate immune responses in HD11 macrophages in vitro and enhanced activities of serum non-specific immune factors. We also identified that APEC\_OMV vaccination could induce both specific antibody responses and a predominant IFN-γ-mediated cellular immune response. Our findings provide broader knowledge to better understand the immunological basis of APEC\_OMV-mediated protection and offer a novel strategy for the development of cross-protective nanovaccines to control various APEC serogroups in the future.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/10/11/2293/s1, Table S1: Primers used for quantitative real-time PCR in this study.

**Author Contributions:** Conceptualization, R.H., Y.G. and M.Y.; methodology, R.H., J.L. and H.L.(Hua Lin); formal analysis, R.H., H.L.(Haojing Liu), M.W. and M.L.; investigation, R.H., J.L., H.L. (Haojing Liu) and M.L.; writing—original draft preparation, R.H.; writing—review and editing, Y.G. and M.Y.; supervision, Y.G. and M.Y. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Natural Sciences Foundation of China (grant numbers 31,672,437 and 31,372,343). The funding organization was not involved in the design of the study, analysis or interpretation of data, or writing of the manuscript.

**Acknowledgments:** We would like to thank the Life Science Core Services facility at Northwest A&F University for microscopic imaging electron microscope analyses.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Primary and Memory Response of Human Monocytes to Vaccines: Role of Nanoparticulate Antigens in Inducing Innate Memory**

**Mayra M. Ferrari Barbosa 1, Alex Issamu Kanno 1, Leonardo Paiva Farias 2, Mariusz Madej 3,†, Gergö Sipos 3, Silverio Sbrana 4, Luigina Romani 5, Diana Boraschi 3,6,\*, Luciana C. C. Leite 1,\* and Paola Italiani 3,6,\***


**Abstract:** Innate immune cells such as monocytes and macrophages are activated in response to microbial and other challenges and mount an inflammatory defensive response. Exposed cells develop the so-called innate memory, which allows them to react differently to a subsequent challenge, aiming at better protection. In this study, using human primary monocytes in vitro, we have assessed the memory-inducing capacity of two antigenic molecules of *Schistosoma mansoni* in soluble form compared to the same molecules coupled to outer membrane vesicles of *Neisseria lactamica*. The results show that particulate challenges are much more efficient than soluble molecules in inducing innate memory, which is measured as the production of inflammatory and anti-inflammatory cytokines (TNFα, IL-6, IL-10). Controls run with LPS from *Klebsiella pneumoniae* compared to the whole bacteria show that while LPS alone has strong memory-inducing capacity, the entire bacteria are more efficient. These data suggest that microbial antigens that are unable to induce innate immune activation can nevertheless participate in innate activation and memory when in a particulate form, which is a notion that supports the use of nanoparticulate antigens in vaccination strategies for achieving adjuvant-like effects of innate activation as well as priming for improved reactivity to future challenges.

**Keywords:** innate immunity; innate memory; *Schistosoma mansoni*; monocytes; macrophages; vaccination

### **1. Introduction**

In vaccine development, antigen-specific immune responses, and the development of long-term protective immunological memory are currently sought by exploiting technological platforms to construct vaccines that allow for appropriate antigen presentation. No vaccines are currently available for the parasite *Schistosoma mansoni*, which is the causative agent of schistosomiasis that is one of the most devastating parasitic diseases in terms of public health and socio-economic impact [1]. Among the most promising *S. mansoni* antigens that are currently considered for vaccine development are the two surface proteins SmCD59.2 and SmTSP-2. SmCD59.2 is a GPI-anchor tegument surface-exposed immunogenic protein [2] orthologue of human CD59 [3], whose function remains to be established [2–4]. At variance with human CD59, SmCD59.2 does not show any activity

A.I.; Farias, L.P.; Madej, M.; Sipos, G.; Sbrana, S.; Romani, L.; Boraschi, D.; Leite, L.C.C.; Italiani, P. Primary and Memory Response of Human Monocytes to Vaccines: Role of Nanoparticulate Antigens in Inducing Innate Memory. *Nanomaterials* **2021**, *11*, 931. https:// doi.org/10.3390/nano11040931

**Citation:** Barbosa, M.M.F.; Kanno,

Academic Editor: Jonathan Lovell

Received: 23 February 2021 Accepted: 1 April 2021 Published: 6 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

on complement [3]. Whether SmCD59.2 could interact with innate receptors, as it has been shown for the human orthologue [5], is currently unknown. SmTSP-2 is an immunogenic tetraspanin protein [2] essential in schistosomula development [4], with structural properties in cell membrane orgnization and in the parasite's tegument [6]. Based on its similarity with human tetraspanins, it is hypothesized that SmTSP-2 could interact with integrins and MHC-II on human cell membranes [4]. Both molecules have been suggested as potential vaccine candidates, although the results to date indicate the need to increase their immunogenicity [7–9]. To this end, these proteins were expressed in recombinant form in fusion with the biotin-binding protein rhizavidin and coupled to biotin-labeled Outer Membrane Vesicles (OMV) of *Neisseria lactamica*, thereby generating antigen-decorated OMV that were more effective in inducing antigen-specific humoral and cellular immunity in mice compared to the soluble protein alone [10,11]. In fact, mice immunized with the antigen-decorated OMV could generate an antigen-specific IgG antibody response much more potent than that induced by the soluble antigen or by the mixture of soluble antigen with bare OMV, the antibody production being paralleled by antigen-specific activation of CD4<sup>+</sup> and CD8+ T lymphocytes in the spleen [10,11].

In assessing the efficacy of vaccine candidates, it is also important to evaluate the effects on innate immunity, i.e., the host response that, although not directly responsible for antigen-specific recognition, reaction, and specific protective memory, has a central role in amplifying antigen-specific responses and in establishing effective long-term immunity. In this perspective, the role of vaccine adjuvants is that of activating innate immunity to improve vaccine efficacy. Similar to adaptive immunity, innate immunity can display memory, i.e., a variation in the secondary response to a challenge, which depends on the host being previously exposed to/primed with the same or other agents [12–16]. At variance with adaptive memory, which is antigen-specific, innate memory is largely non-specific in mammals, with exposure to a given stimulus (e.g., bacterial LPS) causing a secondary memory response that is the same to a variety of different agents [12–16]. Innate memory responses aim at shaping the innate/inflammatory response to secondary challenges in a way that is more protective and less damaging than the first reaction, as in the case of LPS tolerance that limits the extent of the local inflammatory reaction, which would cause significant damage to the affected tissue, while maintaining the production of chemokines and alarmins that initiate the defensive immune reaction [12,17–22]. Most interestingly, vaccination with several whole live attenuated vaccines (*B. pertussis*, BCG, poliovirus, smallpox, measles, measles–mumps–rubella) was shown to induce long-term resistance not only to the specific immunizing microorganism but also to different pathogens [23–26], suggesting that vaccine-induced non-specific innate memory can amplify vaccine-induced protection by extending it to non-related infections. It is notable that particulate agents are very efficient in inducing innate memory, to underline the fact that cells of the innate immune system, in particular mononuclear phagocytes such as monocytes and macrophages, can recognize size/shape in addition to molecular patterns [27].

In this study, we have used an in vitro system based on human primary monocytes to assess the capacity of *S. mansoni* antigens, either soluble or displayed on the OMV surface, to induce innate primary and memory responses, in order to examine the possible contribution of innate immunity in the overall efficacy of the anti-*S. mansoni* candidate vaccines. Our data show that the soluble Sm antigens do not induce significant innate activation of either monocytes or monocyte-derived macrophages in terms of production of the inflammatory cytokines TNFα and IL-6, and of the anti-inflammatory cytokine IL-10, while both bare and antigen-displaying OMV are effective although to different extents. Priming with soluble rSmTSP-2 or rSmCD59.2 did not induce memory to either the same or an unrelated challenge (LPS). Priming with bare or antigen-decorated OMV induced a significant tolerance in terms of TNFα production and a significant IL-10 production in response to LPS but no substantial changes in response to the homologous stimuli. As a control for assessing the role of size in inducing innate memory, the primary and memory TNFα response of human monocytes and macrophages to LPS from *Klebsiella pneumoniae*

was compared to that to whole *K. pneumoniae* bacteria. Although LPS is a potent stimulus of innate responses, also in this case the memory response induced by whole bacteria was more pronounced than that induced by purified LPS.

#### **2. Materials and Methods**

#### *2.1. Synthesis and Characterization of rRzv:SmCD59.2 and rRzv:SmTSP-2 OMV Complexes*

The recombinant fusion proteins between rhizavidin from *Rhizobium etli* and *S. mansoni* CD59.2 and TSP-2 (rRzvSmCD59.2 and rRzvSmTSP-2) were produced in *E. coli*, purified, and characterized as previously described in detail [10,11]. The elimination of possible contaminating LPS was performed with the Triton X-114 wash method [28], yielding recombinant proteins with an LPS contamination <8.0 EU/mg, as evaluated by the LAL gel-clot assay (Lonza Group Ltd., Basel, Switzerland). OMV were obtained from *Neisseria lactamica* 799/98, purified and detoxified by treatment with sodium deoxycholate, and shown to reduce the LPS content by 95% [29]. LPS in membranes is considered 100x less toxic than free LPS [30]; thus, the LPS activity rather than amount was always measured to meet the quality control criteria for OMV vaccines that set the limit of LPS activity to <400 EU/μg. However, the 2-keto-3-deoxy-D-mannooctanoic acid (KDO) measurement (see below) performed on some OMV samples confirmed the presence of an amount of LPS that matched the measured activity, i.e., 1 EU = 0.1 nanogram. Biotinylated OMV were obtained as previously described [10]. Briefly, 25 mg OMV were incubated with 10 mg biotin in sodium phosphate buffer, 150 mM NaCl, 3% sucrose, and 0.1 M *N*-(3-dimetylaminopropyl)-*N*' ethylcarbodiimide hydrochloride, and re-purified by gel filtration chromatography [10]. Recombinant proteins were coupled to biotinylated OMV by exploiting biotin–rhizavidin affinity binding, as previously described for the Multiple Antigen Presenting Strategy (MAPS) [31]. Avidin binding to biotin is the strongest non-covalent interaction known in nature, and it has been extensively developed and approved for many therapeutic applications, including cancer treatments [32,33]. Biotinylated OMV were incubated with rRzvSmCD59.2 or rRzvSmTSP-2 for 18 h at 4 ◦C at a 5:1 mass ratio and then purified by size exclusion chromatography on Sephacryl S-200 in endotoxin-free conditions [10]. The antigen to OMV protein mass ratio after conjugation was 1:10–1:20 [11].

The OMV-protein complexes, displaying SmCD59.2 (OMV:D) or displaying SmTSP-2 (OMV:T) and the unconjugated purified biotinylated OMV (OMV), were characterized by transmission electron microscopy (TEM) using a JEM-1230 microscope (JEOL Ltd., Tokyo, Japan), and for electrophoretic mobility, and for hydrodynamic size/polydispersion by dynamic light scattering (DLS) with a Zetasizer Nano ZS90 instrument (Malvern Panalytical Ltd., Malvern, UK) with a fixed scattering angle of 173◦ at 25 ◦C in triplicate [9]. Endotoxin/LPS contamination was assessed with the gel clot LAL assay (Lonza Group Ltd., Basel, Switzerland) and with the 2-keto-3-deoxy-D-mannooctanoic acid (KDO) assay.

#### *2.2. Human Monocyte Isolation and Differentiation of Monocyte-Derived Macrophages*

Blood was obtained from healthy donors upon informed consent and in agreement with the Declaration of Helsinki. The protocol was approved by the Regional Ethics Committee for Clinical Experimentation of the Tuscany Region (Ethics Committee Register n. 14,914 of 16 May 2019). Monocytes were isolated by CD14 positive selection with magnetic microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) from peripheral blood mononuclear cells (PBMC), obtained by Ficoll–Paque gradient density separation (GE Healthcare, Bio-Sciences AB, Uppsala, Sweden), as previously described in detail [34]. Monocyte preparations used in the experiments were >95% viable and >95% pure (assessed by trypan blue exclusion and cytosmears). Isolated monocytes included the subpopulations of classical, intermediate, and non-classical monocytes at the same percentages as present in PBMC, as indicated by the manufacturer and confirmed in-house by cytofluorimetric analysis of CD14- and CD16-expressing cells (Supplementary Figure S1). The staining procedure and flow cytometric analysis are reported in detail in the Supplementary Materials.

Monocytes were cultured in culture medium (RPMI 1640 + Glutamax-I; GIBCO by Life Technologies, Paisley, UK) supplemented with 50 μg/mL gentamicin sulfate (GIBCO) and 5% heat-inactivated human AB serum (Sigma-Aldrich, Inc., St. Louis, MO, USA). Cells (5–7.5 × 105) were seeded in a final volume of 1.0 mL in wells of 24-well flat bottom plates (well internal diameter 15.6 mm; Corning® Costar®; Corning Inc. Life Sciences, Oneonta, NY, USA) at 37 ◦C in moist air with 5% CO2. Monocyte stimulation was performed after overnight resting.

Freshly isolated monocytes were differentiated into tissue-like macrophages by culturing them in culture medium containing 50 ng/mL macrophage colony-stimulating factor (M-CSF; R&D Systems, Minneapolis, MN, USA) for 6 days (with one medium change on the third day). Differentiation and M2-like polarization, typical of tissue resident macrophages, was assessed morphologically and by the decreased expression of CD14 and increased expression of CD206. No significant mortality or increase in cell number was observed at the end of the differentiation period.

#### *2.3. Human Cell Activation and Induction of Innate Memory*

For assessing the primary response to stimulation, monocytes or macrophages were exposed for 24 h to LPS (positive control; from *E. coli* O55:B5 or *K. pneumoniae*; Sigma-Aldrich, Inc., St. Louis, MO, USA) or to increasing concentrations of rSmCD59.2, rSmTSP-2, unconjugated OMV, OMV:D, OMV:T, heat-killed *K. pneumoniae* (clinical isolates of both wild-type and carbapenemase-producing bacteria), or left untreated (medium/negative control).

For memory experiments, after the first exposure to stimuli for 24 h and supernatant collection, cells were washed and cultured with fresh culture medium for 6 additional days (one medium change after 3 days), to allow for the extinction of the activation induced by the previous stimulation. After this resting phase, the supernatant was collected, and cells were challenged for 24 h with fresh medium alone or containing a ten-fold higher concentration of stimuli. All supernatants (after the first stimulation, after the resting phase and after the challenge phase) were frozen at −20 ◦C for subsequent cytokine analysis. By visual inspection, cell viability and cell number did not substantially change in response to the different treatments.

#### *2.4. Cytokine Analysis*

The levels of the human inflammatory cytokines TNFα and IL-6 and of the antiinflammatory factor IL-10 were assessed by ELISA (R&D Systems), using a Cytation 3 imaging multi-mode reader (BioTek, Winooski, VT, USA).

#### *2.5. Statistical Analysis*

Data were analyzed using the GraphPad Prism6.01 software (GraphPad Inc., La Jolla, CA, USA). For cytokine production, results are presented as ng produced cytokine/10<sup>6</sup> plated monocytes. Results are reported as mean ± SD of values from 2 to 4 replicates from the same donor or from 2 to 4 different donors. Statistical significance of differences is indicated by *p* values, which were calculated using one-way non-parametric ANOVA with *post hoc* Tukey's multiple comparison test and one-tailed unpaired *t* test.

#### **3. Results**

#### *3.1. Particle Characterization*

Unconjugated OMV and OMV–antigen complexes were characterized for their size, polydispersity, presence of recombinant antigens, and LPS content.

The results in Figure 1 show that the three particles have similar characteristics, with an average size between 150 and 250 nm (corresponding to a hydrodynamic size of 200–400 nm) and a negative ζ-potential. All OMV preparations show polydispersity, which is likely due to both particle size heterogeneity and to aggregates [11], as confirmed by TEM. The presence of the recombinant antigens on the surface of OMV:D and OMV:T was confirmed to be about 1:10–1:20 vs. OMV proteins. A more complete characterization of the

OMV, from isolating to the OMV–antigen conjugates, has been published elsewhere [10]. After detoxification, all OMV still displayed a residual LPS activity, which was at least 30 EU/μg in unconjugated OMV, 8 EU/μg in OMV:D, and 80 EU/μg in OMV:T. Since 1 EU roughly corresponds to 100 pg LPS, we can infer the presence of 3 ng LPS per μg particles in unconjugated OMV, 0.8 ng/μg in OMV:D, and 8 ng/μg in OMV:T (the latter value was confirmed by KDO assessment). It should be noted that the OMV complexes comply with the quality control criteria for OMV vaccines that require a residual LPS activity of <400 EU/μg.

**Figure 1. Characterization of Outer Membrane Vesicles (OMV) and OMV–antigen complexes.** TEM images of unconjugated biotinylated OMV of *N. lactamica* (OMV, upper left), OMV conjugated with rRzvSmCD59.2 (OMV:D, upper center) and OMV conjugated with rRzvSmTSP-2 (OMV:T, upper right). Lower left panel: electrophoretic mobility of OMV and OMV–antigen complexes. Left arrows indicate the position of the main *N. lactamica* protein PorB, of SmTSP-2 in OMV:T and of SmCD59.2 in OMV:D (the two antigens having a calculated MW of 27.2 kDa for rRzvSmTSP-2 and 26.9 kDa for rRzvSmCD59.2). Lower right table: Summary of the OMV characteristics of size (measured in TEM, mean ± SD of 16–131 particles), hydrodynamic size, polydispersity, ζ-potential (mean of 3 determinations ± SD) (all measured by dynamic light scattering (DLS)) and LPS activity (measured with the LAL assay). PDI, polydispersity index.

#### *3.2. Innate Response of Innate Immune Cells to S. mansoni Antigens*

The capacity of recombinant *S. mansoni* antigens to stimulate the production of inflammatory (TNFα, IL-6) and anti-inflammatory (IL-10) cytokines was assessed on fresh human blood monocytes and on monocyte-derived macrophage (Figure 2). Both soluble antigens and antigens coupled to OMV were examined in parallel to unconjugated OMV and to the prototypical inflammatory stimulus LPS.

**Figure 2. Primary response of human monocytes and macrophages to** *S. mansoni* **antigens.** The production of TNFα (left panels, round symbols), IL-6 (center panels, square symbols) and IL-10 (right panels, diamond symbols) was assessed in human fresh blood monocytes (upper panels) and monocyte-derived macrophages (lower panels) stimulated for 24 h with increasing concentrations of unconjugated OMV (light blue), rSmCD59.2 (gray), rSmTSP-2 (red), OMV:D (dark blue), or OMV:T (yellow). The negative and positive controls (culture medium alone and LPS 10 ng/mL) are the following: TNFα in monocytes 0.03 and 2.23 ng/106 cells; TNFα in macrophages 0.00 and 3.06; IL-6 in monocytes 0.04 and 5.00; IL-6 in macrophages 0.00 and 1.84; IL-10 in monocytes 0.00 and 0.16; IL-10 in macrophages 0.02 and 0.30. Data are from one donor of 2–4 tested (data from other donors are reported in the Supplementary Table S1). The SD of technical replicates was always <10% and is not reported. Statistical significance is reported in the Supplementary Table S2.

As shown in Figure 2, soluble antigens have little/no activity (gray and red symbols). Conversely, unconjugated OMV (light blue symbols) have a strong capacity of inducing inflammatory cytokines in monocytes and, to a lesser extent, in macrophages, while they are very potent inducers of IL-10 in monocytes and even more in macrophages. OMV:D (dark blue symbols) are more potent than unconjugated OMV in inducing inflammatory cytokines while essentially unable to induce IL-10. On the other hand, OMV:T (yellow symbols) have little/no activity, which is similar to the soluble recombinant protein. A possible interfering role for biotin, present on all OMV preparations for antigen ligation, is likely to be minimal/null, since the three OMV types showed variable qualitative and quantitative differences in the induction of different cytokines while containing the same amount of biotin.

The concentrations indicated in Figure 2 are those of the recombinant proteins, either alone or coupled to OMV; e.g., 1 μg OMV:T is the amount of OMV:T that contains 1 μg of SmTSP-2 (the amount of OMV being about 10× higher, i.e., 10 μg). Likewise, for 1 μg of unconjugated OMV, it is intended that the amount of OMV contained in complexes displaying 1 μg of Sm antigens (again 10 μg). Data in Figure 2 are from one donor, which are representative of two to four tested (the results from all donors are reported in the Supplementary Table S1).

#### *3.3. Innate Memory of Human Monocyte/Macrophages to S. mansoni Antigens*

Monocytes were exposed for 24 h to a low concentration of Sm antigens, OMV, and OMV–antigen complexes (0.1 μg antigen/mL) or LPS as control (1 ng/mL) and cultured for an additional 6 days in fresh medium (one medium change after 3 days) to allow for return to a quiescent state. This was assessed by measuring the release of cytokines in the last 3 day supernatant, which was always undetectable (data not shown). At this time, monocytes had spontaneously differentiated in culture into macrophages. Cells were re-exposed to stimuli for 24 h, the challenge being a 10x higher concentration of LPS (10 ng/mL, as control) or the Sm antigens, OMV, and OMV–antigen complexes (1 μg antigen/mL). The inflammatory cytokine TNFα and the anti-inflammatory cytokine IL-10 were measured. The results in Figure 3 show the memory-induced variation in the cell response to different challenges, which are indicated with different colors. The data in Figure 3 refer to cells from a single donor out of three tested. Given the donor-to-donor quantitative variability of responses, the data could not be averaged (the results from each donor are reported in the Supplementary Table S2). In Figure 3, the response of unprimed cells is reported in the line indicated as "medium" in the horizontal axis "PRIMING", and it shows that cells produce significant amounts of TNFα in response to LPS (orange), OMV (gray), and OMV–antigen complexes (green and dark blue), while the response to soluble rSmCD59.2 (yellow) and rSmTSP-2 (light blue) is limited. For IL-10 production, it is notable that unprimed cells do not respond well to LPS, while the response to OMV is very high. OMV–antigen complexes also induce IL-10 production, although to a lesser extent than bare OMV. This pattern of response reflects quite precisely the response of macrophages to a primary stimulation depicted in Figure 2.

**Figure 3. Secondary response of human monocytes to challenge with** *S. mansoni* **antigens.** Production of TNFα (left panel) and IL-10 (right panel) of human monocytes that had been previously exposed (PRIMING in the horizontal axis) to culture medium alone (medium), LPS (1 ng/mL), unconjugated OMV (OMV), rSmCD59.2 (CD59.2), rSmTSP-2 (TSP-2), OMV:D, or OMV:T (all at 0.1 μg antigen/mL). After 6 days of resting, cells were challenged (see depth axis CHALLENGE) with a 10x higher concentration of stimuli; medium (purple), LPS (orange), OMV (gray), rSmCD59.2 (yellow), rSmTSP-2 (light blue), OMV:D (green) and OMV:T (dark blue). LPS was used as control challenge for cells primed with every kind of stimuli. Data are the values of the 24-h cytokine production by cells from one donor representative of three examined (see Supplementary Table S3 for the values of individual donors). SD of technical replicates were always <10% and are not shown. Statistical significance is reported in the Supplementary Table S4.

When examining the memory response in terms of inflammatory TNFα production (Figure 3, left), we can observe that, as expected, priming with LPS induces a clear tolerance (decrease of response) to an LPS challenge, which is a phenomenon that can be observed also in cells primed with unconjugated OMV and OMV–antigen complexes, whereas priming with soluble Sm antigens slightly increased the secondary response to LPS (see orange columns, LPS challenge). Challenge with unconjugated OMV (gray columns) or with OMV–antigen complexes (green and dark blue columns) showed the same trend, i.e., a decreased TNFα production in cells primed with particulate agents as compared to unprimed cells. Conversely, the soluble Sm antigens triggered in primed cells the same low TNFα production as in unprimed control cells (yellow and blue columns). Thus, our data show that the particulate agents induce a tolerance-type memory in human monocytes/macrophages, which reduces the production of the inflammatory factor TNFα upon a secondary challenge (both identical and unrelated), while the soluble antigens do not have a significant effect. However, it should be noted that OMV also display a significant amount of LPS, roughly corresponding to the LPS control, which may imply that the tolerance effect of priming with OMV could be actually due to LPS.

Quite different is the picture of memory-induced modulation of the anti-inflammatory cytokine IL-10 (Figure 3, right). As already mentioned, challenge with LPS could induce a very limited production of IL-10 in unprimed cells, which was slightly increased in primed cells (except after rSmTSP-2 priming). In the case of priming with particulate agents (OMV, OMV:D, OMV:T), at variance with the results with TNFα, priming did not induce a clear tolerance-type response to challenge with either LPS or the identical agents, with only a partial decrease observed in the case of OMV:T homologous challenge. Thus, the memory response of particle-primed cells results in a significant decrease of the production of the inflammatory factor and no/little decrease of the anti-inflammatory factor, leading to a secondary response that is less inflammatory than that of unprimed cells. In the case of OMV:D, the response of unprimed cells to the complex (18.4 ng TNFα/106 cells and 1.1 ng IL-10) was strongly rebalanced in OMV:D-primed cells (1.2 ng TNFα and 1.4 ng IL-10). In addition, in the case of OMV:T, the response of unprimed cells (10.0 ng TNFα and 1.9 ng IL-10, less inflammatory than the response to OMV:D) was significantly shifted toward anti-inflammation in primed cells (0.2 ng TNFα and 0.7 ng IL-10).

#### *3.4. Innate and Memory Responses of Human Monocytes and Macrophages to Klebsiella pneumoniae LPS vs. Whole Bacteria*

To assess the possible role of size on the capacity of microbial agents to stimulate innate responses and memory, we have tested the production of TNFα by human monocytes and monocyte-derived macrophages in response to LPS from Klebsiella pneumoniae in comparison to the whole inactivated bacteria, which display LPS on their surface. LPS concentrations were selected to correspond to those present on bacteria, considering that 1 EU (corresponding to about 100 pg LPS) is the amount of LPS displayed by 10<sup>5</sup> bacteria. The results in Figure 4 (upper panels) show that the primary response of monocytes to LPS was 7–10× higher than that of macrophages, while the response of macrophages to K. pneumoniae was only half of that of monocytes at the highest concentration.

When examining the secondary memory response, both for LPS and for the whole bacteria, a tolerance-type memory response was evident, with the production of TNFα was much lower in primed vs. unprimed cells (Figure 4, lower panels). The secondary response of monocytes (lower left panel) represents the memory response of effector monocytes that entered an infected tissue and developed memory afterwards, whereas the secondary response of macrophages (lower right panel) represents the memory response of tissue-resident macrophages. The results show that the secondary response of monocytes, which respond to whole bacteria more potently than to isolated LPS, displays a potent tolerance when cells had been primed with whole bacteria, which is a tolerance that is already maximal at the lowest bacterial priming dose (ratio bacteria to monocytes 0.1 to 1). Conversely, tolerance to LPS depends on the LPS priming dose, being minimal at the lowest dose (0.1 ng LPS/106 monocytes, corresponding to a bacteria to monocyte ratio

of 0.1 to 1) and well evident only at a priming ratio of 10 to 1. A similar trend, although less pronounced, can be observed in the secondary response of tissue-like macrophages, in which the tolerance to challenge is significantly more pronounced in macrophages primed with increasing doses of bacteria in comparison to isolated LPS.

**Figure 4.** Primary and secondary innate response of monocytes and macrophages to *Klebsiella pneumoniae* bacteria vs. *K. pneumoniae* LPS. The primary response (upper panels) of monocytes (left panels) and monocyte-derived macrophages (right panels) was measured in terms of TNFα production upon stimulation with increasing concentrations of killed *K. pneumoniae* bacteria (0.1, 1, and 10 bacteria per each monocyte/macrophage; blue columns) and to LPS from *K. pneumoniae* (0.1, 1, and 10 ng/10<sup>6</sup> monocytes/macrophages; orange columns). The secondary response (lower panels) of unprimed control cells (medium, in the horizontal axis; gray columns) and cells primed with increasing concentrations of *K. pneumoniae* bacteria (blue columns) or *K. pneumoniae* LPS (LPS; orange columns) was again measured in terms of TNFα production after a challenge of 24 h with the highest challenge concentration (10 bacteria per monocyte/macrophage; 10 ng LPS/10<sup>6</sup> monocytes/macrophages). Only homologous priming/challenge combinations are shown (in the depth axis). Results are the average of two to four replicates from two different donors. SD were <20% and are not shown. Statistical analysis is reported in the Supplementary Table S5.

It should be said that in these experiments, we have used both wild-type *K. pneumoniae* and carbapenemase-producing *K. pneumoniae* isolates with identical results (the data in Figure 4 are the average values obtained in experiments with different isolates), which suggests that innate immunity and innate memory responses are not affected by the development of antibiotic resistance.

#### **4. Discussion**

The hypothesis that vaccination with a number of bacterial and viral vaccines, in particular those based on live attenuated microorganisms, may increase resistance to a number of unrelated diseases, in addition to the specific disease against which the vaccine is designed, is currently attracting significant attention for the possible positive impact on public health [24]. The biological basis for the non-specific effect of vaccination is most likely residing in the immunological phenomenon known as innate immune memory (or trained immunity), which is well known in plants and invertebrates and also present in vertebrates. Innate memory (also defined as "trained immunity") implies a more efficient innate immune response to a challenge by organisms/cells previously exposed to the same or another agent [13–16,35–37]. In vaccine formulations, the specific immune response to antigens and the development of protective long-term adaptive immunological memory is facilitated by the use of adjuvants, which are agents that non-specifically induce a local inflammatory/innate immune reaction that creates the right conditions for the development of a potent and effective immunization against the vaccine antigens [38]. Indeed, epidemiological evidence supports the hypothesis that recurrent exposure to infectious stimuli is at the basis of a long-term protective immune memory that can be independent of T and B cells, thereby pointing at innate immune cells [39]. Thus, current research on vaccines aims at assessing the capacity of vaccine formulations (in particular their adjuvant components) not only to induce the immediate innate/inflammatory reaction required for optimal specific immunization but also to devise adjuvant strategies able to induce a non-specific innate memory that would increase the host resistance to a wider range of infections/diseases [22–26,40–43]. The concept of vaccination based on innate memory ("trained immunity" vaccines) is being further developed by the notion that organ/tissue-resident innate cells (in particular macrophages) can strongly contribute to innate memory-biased defensive responses to subsequent organ-specific infections [44–48].

In this study, we have investigated the possible role of innate memory, specifically focusing on mononuclear phagocytes (monocytes and macrophages), in the human reactivity to candidate vaccine formulations for *S. mansoni*. In fact, two promising parasite proteins (SmCD59.2 and SmTSP-2) are being developed in particulate formulations that provide excellent immunogenicity in experimental animals, with high production of antigen-specific antibodies and antigen-specific activation of CD4+ and CD8<sup>+</sup> T lymphocytes [10,11]. The vaccine constructs imply the use of OMV from *N. lactamica*, which has been detoxified to decrease their endotoxin content below the acceptable limits for human use, and, using the MAPS technology, biotinylated and conjugated to recombinant fusion proteins encompassing the parasite protein in fusion with rhizavidin [10,11]. The particulate form, which is often used in vaccination strategies as effective antigen carrier, is known to be particularly effective in the interaction with macrophages, which react to particles by proliferating and by readily ingesting the particulate agents, thereby favoring their presentation to T lymphocytes for initiating adaptive immune responses (as macrophages are, together with dendritic cells, efficient antigen-presenting cells) [27,49,50]. Therefore, it is expected that particles, in addition to carrying the vaccine antigens, can have a direct capacity to initiate innate immune responses, at the basis of their adjuvant effect and, consequently, to induce the generation of an innate memory that can contribute to the vaccine efficacy.

The aim of this study was to examine the capacity of the two vaccine formulations for *S. mansoni* not only to activate innate immunity, i.e., to display an intrinsic adjuvant effect, but in particular if and how they can induce an innate immune memory able to contribute to long-term vaccine efficacy and non-specific resistance. To this end, we have

examined the two *S. mansoni* antigens coupled with OMV (OMV:D and OMV:T) compared to unconjugated OMV and to the unconjugated SmCD59.2 and SmTSP-2 recombinant proteins. We should be aware of the fact that detoxified OMV still display endotoxin levels that, although below the threshold for regulatory approval, are detectable and active in our in vitro assays on human monocytes. Specifically, based on its activity the endotoxin levels may be of about 30 ng/mL for OMV, 8 ng/mL for OMV:D, and 80 ng/mL for OMV:T at the highest concentration used for challenge in memory experiments, i.e., 10 μg OMV (corresponding to 0.5–1.0 μg antigen)/mL, while 10x lower in the priming phase. Such endotoxin levels may be at least in part responsible for the capacity of OMV to induce an innate/inflammatory response and to establish a subsequent innate memory. The contribution of LPS in the innate/inflammatory activation of human monocytes and macrophages by OMV was assessed in comparison with similar amounts of purified LPS (from *E. coli*, since LPS from *N. lactamica* was not available; 1 ng/mL as priming and 10 ng/mL as challenge).

The results show different behaviors of OMV in inducing innate immune activation that cannot be exclusively attributed to LPS (see Figure 2). Indeed, OMV:T induced a primary response comparable to that induced by isolated LPS, but it was much lower than that triggered by unconjugated OMV and OMV:D. This is true for all inflammationrelated factors examined, i.e., the inflammatory cytokines TNFα and IL-6 and the antiinflammatory cytokine IL-10. Interestingly, OMV:D are more efficient than unconjugated OMV in inducing the two inflammatory factors, both in monocytes and macrophages, whereas the unconjugated particles are significantly more effective in inducing the antiinflammatory factor IL-10, while LPS and all the other antigens show much lower effects. Thus, the innate reaction to OMV:T is low, in terms of induction of both inflammatory and anti-inflammatory factors, and at similar levels as the response induced by isolated LPS, while the soluble antigens do not induce any significant innate reaction. Interestingly, the unconjugated OMV induced a significant production of inflammatory factors but were also very effective in inducing the anti-inflammatory IL-10, suggesting that their innate immune activation potential and adjuvant effect may be limited. On the other hand, OMV:D are very efficient in inducing the inflammatory factors but not the anti-inflammatory cytokine, suggesting an efficient adjuvant effect. The inflammatory activation induced by the particles turns out to be transient, as expected for a safe adjuvant. Indeed, cell activation is completely extinguished after six additional days in culture, and cells have returned to a baseline quiescent state (again assessed in terms of cytokine production). Thus, we have shown that OMV induce a primary innate/inflammatory activation in both human monocytes and macrophages in culture that goes far beyond their content of LPS, in particular in quantitative terms, with only OMV:T showing effects quantitatively comparable to those of isolated LPS.

To stress the difference between the effect of LPS and that of LPS-bearing particles, we have also compared the innate activation induced in human monocytes and macrophages by different concentrations of LPS from *K. pneumoniae* in comparison to the entire *K. pneumoniae* bacteria displaying corresponding amounts of LPS on their surface (Figure 4, upper panels). In this case as well, bacteria have effects that are not superimposable to those of their LPS, as shown by the dose-independent activation of monocytes by whole bacteria compared to the strongly dose-dependent activation induced by isolated LPS. This causes a lack of response to a low dose of LPS, while the same dose of bacteria induces a significant response, and a very high response to intermediate and high LPS doses, while the response to the corresponding doses of bacteria is essentially identical to that triggered by the low dose and much lower than the response to LPS.

As already mentioned, the hypothesis that a priming of innate immunity could result in the establishment of a longer-term "innate memory" that contributes to an improved secondary reaction to a challenge (an infection, a second vaccine dose) has been recently explored by several groups and proposed as a very promising development in the vaccine field, in particular because of the non-specific effects of innate immunity that may afford

protection against a wide spectrum of infections [22–26,40–45,47]. We have examined the possible role of innate memory upon challenge with the candidate *S. mansoni* vaccines by exploiting an in vitro system based on human primary monocytes that are exposed to the vaccine antigens twice in a priming–extinction–challenge sequence. Essentially, cells exposed in culture to the vaccine antigens for 24 h were subsequently allowed to rest for 6 additional days, so that their primary activation was extinguished and cells were again in a resting state. Then, cells were challenged with the same antigen (or LPS as control), and the response of antigen-primed cells was compared to that of cells that were not previously exposed (unprimed controls). In our study, we have examined the memory response in terms of production of two representative innate cytokines with opposing effects, the inflammatory TNFα and the anti-inflammatory IL-10, in order to have a realistic picture of the secondary innate reaction that includes the balance between inflammation and anti-inflammation. It should be noted that in this study, we have used a single low dose of each compound as priming stimulus, and a higher dose as challenge. This schedule derives from previous studies that have determined it to be the optimal way for assessing priming-induced innate memory, the concept being that memory induced by a lower-level primary stimulation (as in the case of vaccines) can induce a memory effect protective in the case of a strong challenge (e.g., an infection). To assess whether induction of innate memory is an hormesis phenomenon, i.e., whether low vs. high doses of a priming stimulus could afford opposite effects, we had examined the effect of different doses of priming stimuli on the ability to induce innate memory in vitro in human monocytes/macrophages, and we found that the memory effect was the same (tolerance) and dose-dependent, with increasing effect obtained with increasing priming doses, which was a finding that was true for a soluble stimulus, LPS, as well as for a particulate one, zymosan [50].

Examining the memory response in terms of TNFα production, we observed that priming with OMV (either unconjugated or coupled to antigens) results in a tolerance-like response to both a homologous challenge and to LPS. This means that the secondary response of primed cells is significantly lower than that of unprimed controls. This is the same kind of secondary response observed in LPS-primed cells, which reproduces the wellknown phenomenon of LPS tolerance, aiming at preserving the host tissues from damage due to excessive inflammation upon repeated challenges [21,22]. Conversely, priming with the soluble antigens does not induce a different response to homologous challenge when compared to unprimed cells, whereas a potentiation of the response to LPS could be observed. The memory response in terms of production of the anti-inflammatory factor IL-10 shows a very different picture. Challenge with OMV (bare or antigen-decorated) can induce a potent production of IL-10 in control unprimed cells, which is an event that is not mimicked by LPS, showing that the OMV effect is most likely due to their particulate form rather than to their LPS content. Challenge with OMV also triggered a significant production of IL-10 in OMV-primed cells (homologous challenge), which was practically identical to that induced in unprimed cells (except for a partial decrease in OMV:D primed/challenged cells). Thus, while the particle-induced memory results in a tolerance effect for the inflammatory cytokine, the production of the anti-inflammatory cytokine is essentially not affected, suggesting an overall balance toward anti-inflammation in the secondary response. In fact, the TNFα/IL-10 ratios in unprimed vs. primed cells challenged with OMV:D are 16.7 vs. 0.9, and the TNFα/IL-10 ratios in unprimed vs. primed cells in response to OMV:T are 5.3 vs. 0.3, these figures being similar to those obtained with bare OMV (TNFα/IL-10 ratios in unprimed vs. primed cells in response to OMV are 2.2 vs. 0.3). Indeed, even in the case of LPS (a very good adjuvant apart from its toxicity), the TNFα/IL-10 ratio goes toward reduced inflammation in primed cells, being 33.0 (highly inflammatory) in unprimed cells challenged with LPS, and being significantly reduced to 2.5 in LPS-primed cells challenged with LPS. The priming effect of LPS is apparently changed by its presentation to monocytes/macrophages by whole bacteria, as shown in the experiments with *K. pneumoniae*. Monocyte priming with whole bacteria, even at the lowest dose, achieved a very profound tolerance effect (in terms TNFα production), whereas the

tolerance induced by LPS was dose-dependent irrespective of the intensity of the primary response, suggesting that the LPS-bearing whole bacteria induce a more profound memory than the isolated agent.

The present study did not yet address the mechanisms underlying the observed memory effects (studies currently ongoing), which is an issue of particular interest since the observed effects were different (increased vs. decreased production) depending on the cytokine under evaluation, implying concomitant changes at different levels, likely including epigenetic and metabolic regulation [51–55].

#### **5. Conclusions**

These findings indicate that the OMV-based antigens can exploit the adjuvant effects of both the particulate form of OMV and the presence of non-toxic amounts of LPS, for shaping innate immunity and generating innate memory toward the amplification of immune responses with a safe, non-inflammatory long-term innate reactivity. Understanding the mechanisms at the basis of beneficial innate memory would allow us to modulate them in a controlled fashion in order to obtain the desired memory effects in future vaccination strategies as well as in therapeutic immunomodulation. Notably, the innate memory effects appear to be strongly dependent on the donor, which is most likely a consequence of their "immunobiography", i.e., the cumulative effects of their individual history of exposure to repeated and different challenges during the lifetime [56]. This indicates the need for an individual immune memory profile as the basis for a personalized selection of the most appropriate (most effective and safer) preventive and therapeutic strategies.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/nano11040931/s1: Supplementary Materials and Methods (Staining Procedure, Flow Cytometric Analysis); Figure S1, Monocyte subsets before and after positive isolation with CD14 magnetic beads; Table S1, Primary cytokine production by human monocytes in response *to S. mansoni* antigens; Table S2, Statistical analysis of primary cytokine production by human monocytes in response to *S. mansoni* antigens; Table S3, Memory response of human monocytes from different donors primed with *S. mansoni* antigens; Table S4, Statistical analysis of memory responses of human monocytes primed with *S. mansoni* antigens; Table S5, Statistical analysis of primary and memory responses to *K. pneumoniae* bacteria and their LPS.

**Author Contributions:** M.M.F.B. produced the antigens and performed the experiments; A.I.K., L.P.F., M.M., G.S., and S.S. performed the experiments; L.R. provided microbiological reagents and expertise; D.B., L.C.C.L. and P.I. devised the study, analyzed the results, and wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was funded by the European Commission FP7 project HUMUNITY (GA 316383), the Horizon 2020 projects PANDORA (GA 671881), and ENDONANO (GA 812661), the Italian MIUR Flagship InterOmics project MEMORAT, the PRIN project 20173ZECCM, the CNR Italy-Brazil Joint Laboratory initiative, the FAPESP grant 2017/24632-6 (to L.C.C.L.) and Fundação Butantan. M.M.F.B. received a CAPES fellowship.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and the protocol was approved by the Regional Ethics Committee for Clinical Experimentation of the Tuscany Region (Ethics Committee Register n. 14,914 of 16 May 2019). C57BL/6 mice were kept under appropriate conditions during the study according to the Animal Care and Ethics Committee of the institution, under protocol 3314160715.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available in this article and its Supplementary Materials.

**Acknowledgments:** The authors wish to thank Paola Migliorini (University of Pisa, Italy) for her help in the coordination and ethical monitoring of the study, and Richard Malley (Boston Children's Hospital, Harvard, MA, USA) for help with the MAPS technology, including kindly providing the plasmid for expression of proteins in fusion with rhizavidin.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **Cross-Species Comparisons of Nanoparticle Interactions with Innate Immune Systems: A Methodological Review**

**Benjamin J. Swartzwelter 1, Craig Mayall 2, Andi Alijagic 3, Francesco Barbero 4, Eleonora Ferrari 5, Szabolcs Hernadi 6, Sara Michelini 7, Natividad Isabel Navarro Pacheco 8, Alessandra Prinelli 9, Elmer Swart <sup>10</sup> and Manon Auguste 11,\***


**Abstract:** Many components of the innate immune system are evolutionarily conserved and shared across many living organisms, from plants and invertebrates to humans. Therefore, these shared features can allow the comparative study of potentially dangerous substances, such as engineered nanoparticles (NPs). However, differences of methodology and procedure between diverse species and models make comparison of innate immune responses to NPs between organisms difficult in many cases. To this aim, this review provides an overview of suitable methods and assays that can be used to measure NP immune interactions across species in a multidisciplinary approach. The first part of this review describes the main innate immune defense characteristics of the selected models that can be associated to NPs exposure. In the second part, the different modes of exposure to NPs across models (considering isolated cells or whole organisms) and the main endpoints measured are discussed. In this synergistic perspective, we provide an overview of the current state of important cross-disciplinary immunological models to study NP-immune interactions and identify future research needs. As such, this paper could be used as a methodological reference point for future nano-immunosafety studies.

**Keywords:** environmental models; human cells; innate immunity; markers; NPs testing

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. General Introduction: The Need for Studying Nanoparticle–Immune System Interactions**

Over the last twenty years, there has been a significant growth in the research, development, and production of engineered NPs [1]. When materials are downsized to the nanoscale, novel physical and chemical properties emerge, conferring them with new and unique behaviors. Depending on their nature (e.g., composition, size, shape, surface state), these materials have remarkable optical, magnetic, electrical, catalytic, structural, and

**Citation:** Swartzwelter, B.J.;

Mayall, C.; Alijagic, A.; Barbero, F.; Ferrari, E.; Hernadi, S.; Michelini, S.; Navarro Pacheco, N.I.; Prinelli, A.; Swart, E.; et al. Cross-Species Comparisons of Nanoparticle Interactions with Innate Immune Systems: A Methodological Review. *Nanomaterials* **2021**, *11*, 1528. https:// doi.org/10.3390/nano11061528

Academic Editors: Diana Boraschi and David M Brown

Received: 18 February 2021 Accepted: 7 June 2021 Published: 9 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

chemical properties, which can be exploited in many different sectors such as automotive, agricultural, pharmaceutical, and biomedical fields [2–5]. It is estimated that the global nanomaterial production in 2014 was between 0.3 and 1.6 million tons, with SiO2, TiO2 and ZnO nanomaterials being the most abundantly produced [6].

The wide utilization and increasing production of NPs has inevitably lead to an increase in humans and environmental exposure to these materials although exposure routes are not necessarily identical for different organisms. The expected increased exposure in human and environmental organisms has given rise to concerns regarding potential safety risks. The main exposure routes to NPs in both humans and environmental species are highlighted and summarized in Figure 1.

**Figure 1.** The different exposure pathways of engineered NPs that can interact with human or environmental species.

In humans, the first main exposure pathway is via intentional introduction of NPs, for instance during medical administration. The ability of some NPs to interact with molecular and cellular processes and to be target specific makes their use in drug delivery an attractive application. They have long been known to play an effective role in vaccination, acting not only as antigen carriers, but also as adjuvants that activate innate immunity and thereby increase the efficacy of antigen presentation [7]. They can also be valuable tools in medical imaging and diagnosis, and innovative new therapies [8]. Alongside the potential benefits of nanoparticle-based therapies, there is also a risk associated with parenteral introduction of novel substances, and thus there is a need to ensure that NPs will not negatively impact the normal functioning of the immune system [9–13]. Other interactions can arise from passive exposure such as through cosmetic products or food. Although NPs will likely first interact with epithelial and mucosal barriers, in some cases they are able to cross these barriers or potentially cause adverse effects, for example by interacting with the natural gut microbiome [14].

Although most NPs are not directly applied in the environment, many NPs used in consumer products or industry are expected to be released into the environment during production, use or during the disposal of products containing NPs [15]. Over the past decade, an increasing number of products containing NPs have been introduced into agricultural practices with the aim of increasing crop yield and reducing production costs [16]. In addition, the use of wastewater treatment plant biosolids as crop fertilizers can facilitate release of NPs into the terrestrial environment leading to exposure in soil organisms [17]. NPs can also reach aquatic environments, including seashores, through landfill leachates, or direct disposal of wastes (e.g., consumer products containing plastics) [18]. Once in

the water, NPs can remain in suspension in the water column, interacting with planktonic organisms, or due to interactions with organic matter and/or their higher density, NPs can aggregate and deposit on the seafloor. This has been reflected by several models predicting NP concentrations within different regions which showed higher concentrations of NPs in sediments than surface water [19]. Therefore, benthic and sediment dwelling organisms are expected to be exposed to NPs, due to their feeding habit (e.g., filter, deposit feeders) [20]. In addition, some marine invertebrates possess an open (or semiopen) circulatory system, which is in direct contact with the external environment, eventually contributing to increased exposure.

Considering the many possible exposure and entry routes of NPs, defining common parameters for assessing organism-NP interactions is fundamental for allowing comparisons at different taxonomic levels. Innate immunity is a shared feature for every multicellular organism and the effector mechanisms of the innate immune system are the first line of defense that detect and protect the body from nonself objects such as NPs [21–23]. As is the case for natural pathogens, NPs have the potential to induce an immune response. In cases where NPs can elicit an immune response, there is a need to study the type and degree of this response, and the NP-immune interaction mechanisms rather than remaining limited to only measurements of acute toxicity. Comparative immunology, by its multidisciplinary approach may unravel fundamental mechanisms activated by NPs and help further global understanding regarding the effects of NPs.

Experiments in a laboratory are first necessary to allow the understanding of basic mechanisms under controlled conditions. However, carefully chosen models and assessment parameters are important with regard to future translocation to more realistic environmental exposure. To this end, models within this review have been selected which can be good indicators and representatives of their regional and global distribution and which are easy to maintain under laboratory conditions. Environmental models can be therefore compared across taxa and even to human cells, through both in vitro and/or in vivo approaches according to the model possibilities (Figure 2). Plant models are a compelling place to begin for assessment of NP-immune interactions. In particular *Arabidopsis thaliana*, a small flowering plant belonging to the Brassicaceae family, which is widely used in crop science studies and was also the first plant genome to be fully sequenced [24–26]. Among terrestrial invertebrates, earthworms belonging to the family Lumbricidae *(Eisenia fetida)* are abundant in the soil and play an essential role in soil formation, by facilitating nutrient cycling, fragmenting biomass and aeration of soil through bioturbation [27,28]. Similarly, terrestrial isopods, such as *Porcelio scaber*, are crustaceans which evolved to live on land, inhabiting the top-soil level. They are decomposers and play an important role in returning nutrients to the soil [29–31]. Their feeding habits makes it likely they will come into contact with environmental pollutants, including NPs, and thus represent interesting model species to study these interactions. The Mediterranean mussel *Mytilus galloprovincialis* and the sea urchin *Paracentrotus lividus* are both sessile marine invertebrates. Mussels are able to filter large quantities of water which they use for breathing and feeding, while sea urchins graze on the seafloor layer. These qualities, as well as the ease with which they can be harvested along seashores, make these good models in which to study invertebrate interactions with NPs [20,32–34].

This work is supported by the EU PANDORA project [35], which devoted effort to study the effects and mechanisms of action of NPs on the innate immunity of different models from across the tree of life. The general outcomes of the project were previously reported, summarizing the main findings but also to set future perspective and research direction in this field [21,36]. The remainder of this review will focus on the translatable aspects of experimental methodology, parameters and endpoints used, the suitably of the selected models when considering investigating NP effects, and the possibilities regarding research at the whole organism level (in vivo) or with isolated cells (in vitro).

**Figure 2.** The different models discussed in the current review, and their main experimental usage in the laboratory: in vivo (whole organism experiments) and in vitro (isolated cells or cell lines).

Here we aim to: (i) give a short overview of the characteristics of various relevant innate immune models from across tree of life; and (ii) provide a comparative analysis of the methods used to study the interaction of NPs with these innate immune models.

#### **2. Short Description of the Innate Immune System for the Models of Interest**

#### *2.1. Generalities and Conserved Innate Immune Traits across the Selected Models*

The ability to mount an immune response against external threats is a characteristic of every living organism. While increasing levels of immune complexity are found in higher organisms, at every stage of evolution there is present a basic initial host defense that has been characterized as innate immunity. Innate immunity is a fast, standardized, nonspecific response which includes multiple levels of defense mechanisms, beginning with physicochemical barriers (e.g., shell, mucosal or epithelial barrier) [37,38]. Further mechanisms of defense rely on dual components of the immune system, the immune cells (e.g., monocytes, macrophages for vertebrates, or hemocytes, coelomocytes for invertebrates) and the production of humoral factors. Innate immune cells found in the circulating fluids of invertebrates can have different names and the cellular portion of their innate immunity relies on these unique cells. These cells can be subdivided into different cell populations, such as granular or hyaline cells, and they have distinct roles and can trigger a specific response upon encountering threatening nonself material. Only plants lack these specialized immune cells, but in plants all the cells are believed to be able to mount a defensive response to foreign attack [39]. Complex machinery, including cells and humoral factors is involved in recognition of nonself material, and especially in detecting domains called pathogen/microbe associated molecular patterns (PAMPs/MAMPs) that are typically displayed on the surface of bacterial, fungal, and parasitic organisms and virus-infected cells. Host recognition of nonself will involve a large range of cell membrane bound and scattered pattern recognition receptors (PRRs). Although the distinctive PRRs can vary between models, the main concept is consistent, and different PRRs share a similar role upon recognition of nonself particles. PRRs in humans, much like their invertebrate homologues, are responsible for initiating innate immune cell responses including the initiation of phagocytosis or endocytosis, cellular motility, and beginning the processes leading to inflammatory reactions [40]. Upon successful recognition, the pathogen detecting cell will initiate a process of destruction or sequestration to eliminate eventual danger, and later repair the stress or damage caused by this unexpected material. Most invertebrate immune cells, similarly to human macrophages or monocytes, are involved in phagocytosis, which remains one of the most efficient mechanisms to clear nonself material. The induction of some global defense mechanisms can be easily observed across different models, such as the production of reactive oxygen species (ROS) and nitrogen radicals (RNS), synthesis and secretion of antibacterial and antifungal proteins, cytokine-like proteins, and hydrolytic enzymes. Antimicrobial peptides (AMPs—small cationic, amphipathic molecules) are very well studied and highly involved in invertebrate immunity. They can tag objects or induce direct destruction by destabilizing biological membranes, which make them effective against large range of unicellular organisms like bacteria, yeast, fungi, and also

some protozoans and enveloped viruses [41]. Circulating fluid also contains a large panel of enzymes (released by immune cells) with hemolytic, proteolytic and cytotoxic roles (one of the most common being lysozyme).

The encapsulation of foreign objects and activation of enzymatic cascades that regulate melanization and coagulation of hemolymph are also common defense mechanisms encountered in invertebrates. Indeed, phenoloxidase is considered among the most important components of the invertebrate immune system, especially in insects and crustaceans. The phenoloxidase cascade produces the antimicrobial molecule melanin, as well as inducing multiple potent bioactive agents such as peroxinectin and ROS, that aid in phagocytosis and cell adhesion. Melanization is essential in wound healing, encapsulation, and nodulation. Proper modulation of this enzyme is crucial to ensure survival of the organism. The majority of species activate the phenoloxidase cascade using the proPO enzyme [42–44].

Although general immune features are conserved across the previously described models and organisms, adaptations exist in each case that address the organism particular vulnerabilities and environmental stresses. These adaptations occur according to the organism's lifestyle, habits and need, which might cause certain parameters to be more important than others in some species to deal with threats some models are more likely to encounter. In line with this, the next section aims to report the main mechanisms and characteristics of innate immune responses for the selected models, and in particular those known to be activated upon exposure to NPs. A summary is presented in Table 1.


**Table 1.** Summary of the main defense mechanisms involved in innate immunity at different levels of the models discussed.


**Table 1.** *Cont.*

<sup>1</sup> Other innate cell types exist that are not discussed, including natural killer cells and innate lymphoid cells. Refer to the main text for the meaning of the abbreviations.

#### *2.2. Model Specific Immune System Characteristics*

#### 2.2.1. Plants

Plants lack specialized mobile immune cells, and every cell is believed to be capable of initiating an immune defense against pathogens and invaders. Two layers of innate immune responses, i.e., pattern triggered immunity (PTI) and effector-triggered immunity (ETI), provide an efficient defense that keeps most pathogens and external attacks under control [39,45]. The activation of PTI relies on PRRs perceptions of MAMPs/PAMPs to trigger complex immune responses. PRRs are solely on the cell surface of plant cells and among them the most studied is the plant flagellin receptor-FLS2 [46–51]. The second line of defense, ETI, is mediated by nucleotide-binding domain leucine-rich repeat (NB-LRR) disease resistance proteins (NLR), which induce defense responses leading to a hypersensitive programmed cell death [52]. NLRs can detect effectors directly or by indirect surveillance of the effector action on other host target proteins [39]. Both lines of defense share parts of their defense signaling pathways [53,54]. After pathogen detection, multiple morphological and physiological responses are induced, such as ion fluxes over the plasma membrane, including Ca2+- and H+-influx; production of ROS and antimicrobial compounds (phytoalexins); activation of mitogen-activated protein kinases (MAPKs) and calcium-dependent protein kinases (CDPKs). In consequence, the transcriptome will be reprogramed by activation of a subset of transcription factors; callose deposition; stomatal closure; restriction of nutrient transfer from the cytosol to the apoplast and programmed cell death [55,56]. Phytohormones such as salicylic acid, abscisic acid, jasmonic acid, and ethylene have a critical role in the plant's responses to specific pathogens [57]. Defense hormones can be transported within and between plants to alert distant tissues and confer systemic immunity [55,58].

#### 2.2.2. Earthworms

Earthworms are protostome animals that have large coelomic cavities throughout the length of the animal. The coelomic cavity is typically nonsterile, open to the outer environment through dorsal pores, allowing the entrance of fungi, bacteria, and protozoans. Coelomocytes can be classified into two major cell types: amoebocytes and eleocytes. Amoebocytes (hyaline and granular) are involved in various immune responses including phagocytosis, encapsulation, and the production of antimicrobial molecules [59,60]. Eleocytes display more nutritive and accessory functions [59,61]. Three types of PRRs have so far been identified: coelomic cytolytic factor (CCF) [62], toll-like receptors (TLR) [63], and lipopolysaccharide-binding protein/bacterial permeability-increasing protein (LBP/BPI) [64]. CCF has two recognition domains that can interact with bacterial or fungal MAMPs, which in turn triggers the proPO cascade [42,65]. A range of antimicrobial molecules including

lysozyme and the hemolytic proteins fetidins and lysenins are involved in the elimination of the microorganisms [66–69].

#### 2.2.3. Isopods

The hemocytes of *Porcellio scaber* originate in the hematopoietic glands located along the animals dorsal vessel, and can be split into granular and hyaline hemocytes [70,71]. Hyaline (absence of granules) hemocytes are mainly responsible for phagocytosis [71]. Semigranular cells also show some phagocytic ability but seem more involved in encapsulation and nodulation. Granular cells are predominantly involved with the phenoloxidase system [72], and along with semigranular cells are thought to produce AMPs and be involved in antioxidant defense [73,74]. In *P. scaber*, the PO cascade is initiated by hemocyanin [44]. In addition, for defense they are able to produce, RNS, ROS, and AMPs like other invertebrates [75,76]. Genomic mining of the terrestrial isopod, *Armadillidium vulgare*, revealed genes for specific AMPs including anti-lipopolysaccharide factor (ALF) 1 and 2, crustin 1, 2, and 3, and I type lysozyme, and pathogen recognition genes C-type lectins 1, 2, and 3 and peroxinectin-like A and B [77].

#### 2.2.4. Mussels

The innate effector cells of *Mytilus*, hemocytes, are composed of granulocytes and hyalinocytes. Mature granulocytes are among the first lines of cell defense for the elimination of invaders via phagocytic processes [78]. In *Mytilus*, a large range of PRRs, anchored on the cell outer membrane and secreted are encountered, with lectins the most dominant group. Other classes of soluble PRRs are found, like C-terminal fibrinogen related domain-FReD-containing proteins, which have been shown to improve the rate of phagocytosis. TLRs and peptidoglycan recognition proteins-PGRPs, and others have been recently discovered but further study remains to be done to properly appraise their mechanisms of action (see [79] for more details). Moreover, *Mytilus* possesses the complement system pathway and relies on the involvement of C1qDC (C1q domain-containing) proteins [80]. Several signaling transduction pathways have been reported to be present in bivalves such as the mitogen-activated protein kinase (MAPK), nuclear factor-κB (NF-κB), the complement component, the toll pathways, and the JAK-STAT pathway (reviewed in [79]). Additionally, as with other species, hemocytes can trigger the production and release of several factors such as ROS, nitric oxide-NO, hydrolytic enzymes (e.g., lysozyme), and AMPs. Several AMPs have been identified in *Mytilus* such as mytilin, myticin, mytimicin (with an antifungal and/or antibacterial role) and defensins [81]. In extreme cases and for larger objects, hemocytes can encapsulate foreign matter via the coordination of several hemocytes and the release of cytotoxic products (enzymes or ROS) to degrade the material, followed by cellular reabsorption of the debris [82]. Finally, the proPO cascade, while present, remains relatively unstudied in bivalves [83].

#### 2.2.5. Sea Urchins

Sea urchins contain circulating immune cells called coelomocytes which can be subdivided into four classes and are able to infiltrate into different tissues [84,85]. The macrophage-like phagocytes can encapsulate and internalize nonself particles, red amoebocytes release the bactericidal pigment echinochrome A, and white/colorless amoebocytes operate the cytotoxic/cytolytic response, while vibratile cells most probably degranulate and trigger immune cell aggregation [85]. Recently, their genome sequences have revealed the presence of a vast array of immune-related genes, including those coding for PRRs such as TLRs, NLRs or SRCR domain-containing proteins, and complement proteins (Complement C3 homologue) [86,87]. Moreover, lectins are also important in sea urchins and in addition to their role in opsonization, they show lytic functions, and are involved in wound repair [88]. Sea urchins contain several humoral factors including hemolysin and agglutinin which can be induced upon cell activation. ROS are also produced during immune responses. Echinoderms possess many different AMPs with various modes of action

depending on the species, among them paracentrin 1 in *P. lividus*, showing an antimicrobial role [89]. Interestingly, the phagocytes can contain AMPs but they are not released into the extracellular medium, instead they play a role within the phagolysosome [88]. Finally, sea urchins belong to deuterostome lineage which makes them phylogenetically near to chordates, sharing several common traits with mammalians, especially with regard to cytokine production [90,91].

#### 2.2.6. Human Cells

Human and other mammalian immune responses are organized within two branches: the innate immune response which is characteristic of all eukaryotes, and additionally a highly specific adaptive immune response, individualized for distinct pathogens. As NPs do not display highly specified and unique surface patterns, it is the innate branch of human immunity that is tasked with responses to NP exposure. Nonself particles and pathogenic threats that enter human circulation may activate humoral components such as AMPs and complement alongside innate immune cells. Cells participating in the human innate immune response include granulocytes, such as neutrophils (which primarily function to overwhelm pathogenic invaders through large numbers and phagocytic mechanisms), and myeloid-derived cells, including monocytes, macrophages, and dendritic cells (DCs). Monocytes represent about 2–8% of the leukocytes in circulation at any given time [92] and generally patrol the circulatory system for signs of foreign particles or internal damage. They can be recruited in tissue via resident cells releasing chemokines such as CCL2 [93]. As monocytes attempt to engulf foreign particles by phagocytosis, simultaneous chemokines and cytokines are secreted that signal for a broader inflammatory and immune response. They can be further involved in the resolution of an inflammatory reaction, assisting in tissue repair [94]. Macrophages, are tissue resident and represent up to 15% of the cells in a given tissue [95]. Functionally two broad classes of macrophage exist, M1 which display a more inflammatory phenotype involved in early immune response (killing and defending); and later M2, which display more phagocytic and tissue repair oriented traits [96–98]. Finally, DCs are known as antigen-presenting cells, acting as the bridge between human innate and adaptive immunity. They play a role in the generation of pathogen-specific T-cells and B-cell antibodies. Of the PRRs in humans, toll-like receptors (TLRs) play the most prominent role in the detection of extracellular pathogens [99], where they recognize substances such as bacterially associated carbohydrate patterns or RNA sequences associated with viruses [100]. Other PRRs that can be found on the membrane of human innate immune cells include scavenger receptors, which detect various polymers and lipoproteins [101], and C-type lectin receptors including dectin-1, which recognizes B-glucan components of various fungi [102]. In addition, they are some intracellular PRRs found in cytosol, such as NOD-like receptors (NLRs) and rig-I-like receptors (RLRs), which recognize a large range of PAMPs [103–105]. The most notable difference with invertebrates, is the diversity and number of types of cells involved in the immune response.

#### **3. Parameters Assessed: From NPs to Innate Immune Responses**

#### *3.1. NPs: What to Consider When You Use a Biological System?*

The physico-chemical characteristics of a NP and its behavior in different exposure media are fundamental considerations when attempting to understand the interactions of NPs within a biological model. It is important to take into account that the relatively large surface area to volume, the low coordination of atoms at the surface, and their colloidal nature cause NPs to display physical and chemical characteristics that differ from their bulk counterparts. It is also fundamental to understand the characteristics of the final object that living organisms will encounter and to correlate the pristine and final NP features with the potential effects on living organisms. The main NP characteristics to be considered are reported in Figure 3.

**Figure 3.** The different characteristics of NPs and parameters to investigate when they are in suspension media for laboratory experiments.

#### 3.1.1. Primary Characterization

The first determination of primary characteristics includes the description of the material composition, the nominal size, shape, and surface charge (zeta potential). Moreover, characterization of NP coating and other surface modifications are crucial to consider (Figure 3 left panel).

#### 3.1.2. Behavior in Medium

In addition to the known properties of a chosen NP following synthesis, once exposed to biological conditions (e.g., medium or circulating fluid), NPs can display unpredicted new characteristics. NPs can have a propensity to move towards a more stable thermodynamic state via different means: aggregation (which can mean escaping from the nanoscale), formation of a coating composed of various molecules, chemical transformations, particle corrosion, and dissolution [1]. All these transformations can change the identity of the NP or produce new chemical entities (e.g., reactive metal ions), modifying their behavior and their potential associated risk and interactions (Figure 3 right panel). Therefore, the determination of NP characteristics in exposure medium needs to be assessed. This generally includes the aggregation state (Z-average), the change in surface charge (zeta potential), and the dispersion index (PdI). Moreover, the evolution over time of these parameters can also be of value for a full appraisal of the NPs dynamic in the exposure medium. All these analyses are usually performed using DLS (dynamic light scattering) analysis, or electron microscopy (TEM and SEM) depending on the material being investigated. Additionally, careful controls have to be performed in order to avoid artifacts due to the presence of chemicals, often used to stabilize the particles [106,107], or contaminants, such as bacterial lipopolysaccharide (LPS), that can cause false positives in an immune assay [108].

Another routinely measured parameter is the presence and composition of the molecular biocorona, where components of biological fluids can be adsorbed by the NP, forming a corona on its surface. Usually, they are believed to be mostly constituted by proteins (protein corona -PC) but other macromolecules including lipids present in the medium can also contribute to its formation. The presence of this supplementary layer on top of the NPs can in turn affect the NPs behavior and interactions with the surrounding media. However, this corona depends on both the biological fluid (plasma, or otherwise) composition and the properties of the NPs, including size, curvature, surface functionalization, and charge. The composition of the corona is theoretically divided into the soft corona (weakly bound) and the hard corona (tightly bound), but it is dynamic and the ligand on the top can be

exchanged and replaced over time, according to the affinity of the macromolecule for the NPs [109,110]. The PC is the biological identity of the NP and represents what cells "see" and with which they will interact [111–118]. Consequently, recognition by immune cells can be different and specific from one type of NP to another, which means that they will interact with the protein on the surface rather than the NP itself. This results in the triggering of defense mechanisms different from those observed in medium free of proteins. This does not apply only to mammalian plasma, but it has been demonstrated in the biological fluids of different terrestrial and marine invertebrates, including earthworms, bivalve, and sea urchins, in which the composition and effects on immune parameters appeared different for each NP type [119–122]. For these reasons, the PC is an important parameter to consider under laboratory conditions and needs to be characterized with precision during the exposure event.

All the previously cited characteristics can also be applied to environmental media [123–126]. The NPs will be subjected to other factors like abiotic physico-chemical parameters (such as pH, ionic strength, temperature) which can influence their dispersion, aggregation, agglomeration [127,128], interaction with molecules present in their environment, and adsorption to macro-organic matter (e.g., eco-corona) [117]. Scientific literature is being produced on the physico-chemical transformation of NPs due to their exposure to aquatic and terrestrial scenarios, correlating the environments and particle properties with the observed changes. Consideration of this should be taken into account in future studies working with environmental scenario experiments [129–132].

#### *3.2. Models, Cell Culture and Mode of Exposure*

In experimental science, the use of in vitro assays is being promoted as sustainable alternative for a large range of product testing, including NPs, following the 3R principle (replacement, reduction, and refinement). The extraction of immune cells, separation, culture and feasibility to maintain such isolated primary cells varies across models. To illustrate, a summary of the different methods of cell harvesting and exposure for both invertebrates and human primary cells are reported in Figures 4 and 5.

**Figure 4.** The different in vitro approaches and NPs exposure parameters encountered across the selected models.

#### 3.2.1. Nonmammalian In Vitro Assays

In many invertebrates, coelomocytes/hemocytes act as the first line of defense against nonself objects. The induction of functional responses with these cells is often rapidly observed, helping to counteract the limitation of the relative short-term lifespan of cells in cultures (ranging from a few hours to a few days depending on the model). As a natural defense mechanism, earthworms can extrude their coelomic fluid, through dorsal pores. Therefore, coelomocytes can be extracted by mild electrical stimulation or by exposing the animals to an irritative substance [133,134]. Immediately upon collection, the coelomocytes need to be stabilized in a culture medium in order preserve cell viability. Recent studies show that RPMI 1640 medium is the optimal medium for earthworm coelomocytes culturing, as well as the assessment of NP toxicity towards coelomocytes [135,136]. Critical for the successful culturing of coelomocytes is the adjustment of osmolality of the medium so that it reflects that of the coelomic fluid [137–139]. Exposure time for in vitro assays depends on cell viability and may range from between 2 to 72 h, with 24 h being the optimum time for cell cultures in RPMI 1640, according to some investigations [136,137,140]. For terrestrial isopods, the culturing of hemocytes did not show hopeful results yet. Hemolymph can be collected by puncturing through the intersegmental membrane on the dorsal side of the isopod with a sterile needle and collecting the hemolymph with a micropipette. With the use of ringers solution and a MAS (mitochondrial assay solution) buffer, cells appeared to hardly survive for even a few hours outside the body. The selection of a suitable medium is still needed to be identified and adapted for keeping hemocytes alive without showing excessive levels of stress [141].

In marine invertebrates, in vitro experiments are much more abundant, in particular, experiments using hemocytes of the marine mussel *M. galloprovincialis* extracted via a noninvasive method. This method can provide a first line of investigation for testing several types of substances, including NPs [142–145]. *Mytilus* hemolymph is easy to collect via the adductor muscle and fluid quantities are sufficient (depending on the season, volume can be as high as several ml per animal) to perform various experiments [146]. Short-term exposures (≤1 h) have shown rapid activation of hemocyte functional parameters but longer exposure times (up to 24 h) have shown the induction of further immune or stress parameters. For short experiments, hemocytes can be maintained in a natural hemolymph or seawater suspension, in tubes or as monolayers on glass slides. Longer culture times were more successful when modified synthetic basal medium (Basal Medium Eagle) was used in microwell plates. These in vitro experiments are possible due to the cells ability to quickly adhere to supports (<20 min) [147,148]. Ex-vivo tissue explant has also been used (e.g., gills) to study the first interactions and potential uptake of NPs [149].

The coelomocytes from sea urchin once extracted are placed in cell culture plates and kept in EGTA-containing cell culture medium and artificial seawater [84,85,150]. The coelomocytes can be kept for a long period of time (over two weeks), with regular medium replenishment and without the addition of the special growth factors or nutrients [90,151].

Finally, as plants do not possess specialized mobile immune cells, in vitro research is not typically suitable/realizable, and the main experiments in laboratories are made on the full plant or tissues excisions (see next section). Each piece of tissue should respond upon exposure as each single cell is able to launch an effective immune response [152].

#### 3.2.2. Human Cell Models

The study of human cells offers a wide range of possibilities not currently developed for invertebrate models. Multiple cell types, coculture conditions, and cell maturation or differentiation programs exist to define more precisely the interactions with NPs. Several important models used in mammalian systems to test NP–immune system interactions are listed below.

**Figure 5.** The different in vitro NP exposure possibilities available for human primary cells and the procedure of cell extraction and preparation (Reprinted with permission from Michelini et al. (2021) [153]. Copyright 2021, Copyright Royal Society of Chemistry).

In vitro modelling of human innate immunity is usually conducted using monocytes, macrophages, or dendritic cells, as these cells are responsible for directing the innate immune response from pathogen recognition to phagocytosis to inflammation, and even to eventual antigen presentation and induction of adaptive immunity. Cell lines for each of these cell types exist and are frequently used due to their easy experimental repeatability and scalability, with THP-1 (monocytes) and U937 (macrophages) being the most frequently reported [154]. However, cell lines are truly limited to the representative phenotype observed at the time of culture, and even this is susceptible to mutations that do not represent the true reactivity of healthy human cells. More robust models of innate immune responsiveness utilize primary cells, which are collected directly from donors and may be isolated using techniques that select for the desired cell type. Primary cells are representative of an individual's current in vivo condition, and lack the altered metabolic and epigenetic profile inherent to cell lines. Furthermore, as monocytes are found abundantly in circulation, and since they can be precursors for both macrophages and DCs, the differentiation of monocytes in culture into primary differentiated macrophages or DCs is an effective tool to create models of innate immune responses in vitro. However, models utilizing primary cells must contend with individual variability as the immune experience and capacity is different between donors [155].

The whole blood assay is one of the most simple and rapid tests for assessing the immune activating capacity of novel substances within a human system. Typically, blood is drawn from healthy donors and immediately exposed to the substance under investigation, with 250 μL of blood typically diluted in 750 μL of RPMI plus the tested material and incubated for 24–48 h [156]. Peripheral blood mononuclear cell (PBMCs) can also be

isolated from the whole blood using Ficoll–Paque density gradient centrifugation [157]. Magnetic cell separation using some CD-4 beads can be used to isolate monocytes, and later growth factors can be added to differentiate macrophages such as macrophage colony stimulating factor (M-CSF) or DC GM-CSF and IL-4 [94,158].

The monocyte activation test (MAT) models (or using macrophage and DCs) can assess the exposure and response of monocytes to NPs, and many parameters may be assessed following activation [159–161]. Usually cells (in the range of ~500,000 cells/mL) in plate culture can be directly exposed to NPs added to the wells, and tests on monocyte/macrophage/DC activation are typically completed within 24 h. Oftentimes, PBMC culture is conducted in round bottom wells, which simulate a lymph node in which communication between myeloid cells and lymphocytes occurs.

Finally, NPs can also interact with other cells present in blood, such as DCs, which link with adaptive immunity. Similar principles of the MAT test can also be applied for testing NPs, but also cocultures with T-cells of self or foreign origin [153]. These types of test can mimic autoimmunity or the mixed lymphocytes reaction and are of interest for the use of NPs in vaccines and immunotherapy [162].

Experiments considering primary isolated cells offer other advantages by representing a simplified model, limiting interfering factors, which could help to spotlight NPs mechanisms of action before performing further experiments; e.g., coculture, tissue models, or even using whole-organism in vivo experiments. Moreover, these in vitro experimentations allow easier comparison between models, particularly, with human cells. In addition to the commonly known pros for in vitro assessment such as cheap cost, fewer animals used and relatively fast results, a list of the more important pros and cons for each model, with special input for in vitro assays, is presented in Figures 4 and 5, lower panel. Although they provide a simplified set up and can be used to try and understand some of the basic mechanisms, they do not represent the true exposure pathway. Additionally, large variations in the exposure time and the culture methods between different models persist. In invertebrates, the immune cells are usually easy to collect, except for isopods, and in large quantity. There are some species-specific difficulties in experimentation, such as molt cycles in isopods or seasonality with reproductive period in mussels that can impact immune measurements. Moreover, some cells are more sensitive and fragile to handle compared to others. Usually, the immune cells from invertebrates are viable for shorter times in culture, as basal parameters are quickly impacted. For humans, in addition to regulatory hurdles, donor availability is restricted for the obtaining of primary immune cells. Each donor is usually considered independently, which can reflect stronger variabilities in responses. However, from one whole blood sample many cells can be collected and offer a large range of possible assays after purification.

#### *3.3. Whole Model Exposure Experiments*

In vivo experiments allow for evaluation of the effects and mechanisms of action of NPs in organisms at different levels of biological organization (molecular, cellular, tissue level). They provide a realistic scenario of the exposure pathways as encountered under natural conditions. For controlled laboratory experiments, the mode of exposure to NPs needs to be adapted for each model; a summary is presented in Figure 6.

These tests are usually conducted in environmentally relevant mediums (soil or water), through feeding experiments or through breathing and filtering experiments for aquatic species. As the selected models are usually easy to maintain in laboratory conditions for long periods of time, requiring little space and maintenance (e.g., feeding), the exposure time (acute, semichronic and chronic exposure) can range from hours, to days and weeks.

**Figure 6.** NP exposure approaches using the whole organism with the different exposure pathways across the selected models.

Plants can germinate and grow directly in the presence of the NP in the growth medium, i.e., soil, hydroponic nutrient solutions or agar-solidified agents, or they can be exposed at subsequent development stages. Despite the presence of cell walls, that can represent a barrier preventing NPs entering into the plant cell and cytoplasm, NPs might be absorbed through root or leaf and be potentially transported to the shoot or to other points through the phloem (vascular system) [163–166]. NPs can be also dispensed onto the plants surface by foliar spray application [167]. After entrance into the leaf tissue, NPs can diffuse into the intercellular space, the apoplast, or membranes and cause secondary effects. Moreover, temporality is important, and it is necessary to understand the course of plant growth and development, from seed germination to root elongation and shoot emergence, in relation to NP exposure [168]. The following investigations can assess the NP uptake by cells and further nanophytotoxicity, focusing on the toxicity symptoms of plants.

In vivo earthworm exposures are typically conducted in soils following well-described and standardized procedures (e.g., [27]) that can also be applied for NPs [140,169]. However, care must be taken when it comes to the mixing of NPs with soils, adjusting the parameters depending on the form (i.e., as solution dispersion or as powder) in which the NPs are supplied [170]. Furthermore, coexposures with infectious microbes are also important in order to establish whether an exposure to NPs has an effect on the ability of a host to maintain immunity [171]. A methodological approach to investigate the impact NPs have on the earthworm's ability to maintain immunity when coexposed with infectious bacteria has been recently established [140].

A major benefit of working with the terrestrial isopod, *P. scaber* is that they are able to be exposed to the NPs in a manner similar to how they would be exposed in nature. NP suspensions can be spread on leaves that *P. scaber* eat, and both the leaf and animal are then placed in a petri dish. During the experiment, feeding rate, defecation rate, and mortality can be monitored. This also allows for modelling of real-world impacts of NPs on the organism from behavioral changes, like feeding avoidance and mortality,

to cellular immune responses. However, the gut of *P. scaber* is covered in a thick cuticle which is believed to stop the translocation of NPs from the gut into the hemocoel where the hemocytes are, so the immunological effects of the NPs might not be seen when ingested [172]. There is the possibility for an alternative exposure scenario, with injection experiments delivering substances directly into the animals hemocoel, allowing for the study of the direct interaction of a known concentration of NPs with the hemocytes. This ensures NP and immune cell interaction [141,173].

Mussels are suspension feeders and are able to filter large amounts of water (up to 3 L per hour) implying that, in a short period of time, they can easily uptake the NPs present in the seawater of experiment tanks. For this reason, the NPs can be directly added to the seawater and the ventilation system allows for constant movement of water within the tank. To study the first immune defense response, short term experiments (24 h to 96 h), have been shown to be sufficient to induce the activation of the immune system [142]. Moreover, the use of artificial seawater (ASW) implies the absence of organic matter or other substances that could interact with the NP suspension; together with a constant salt content and as such are reproducible for all periods of the year. The experiments are mainly conducted in the spring and summer periods where mussels are at their healthiest. During experiments, mussels are not fed and can readily survive several days without feeding [146]. This is necessary for NP experiments, as the presence of microalgae could interact with the NPs. In this context, the study relies only on the uptake of the NPs in seawater. To mimic a more realistic exposure, other studies have performed longerperiod experiments to consider the interactions between food intake and the NPs, but this generally focused more on physiology and tissue changes and not strictly the immune response [148,174]. Biological uptake routes are dependent on NP properties and may occur as direct uptake in gill tissues and/or through transference from the cilia to the digestive system. Moreover, the agglomeration of particles in seawater has been shown to facilitate NP ingestion by suspension feeding bivalves, and their potential translocation from the gut to the circulatory system [175,176]. However, this internalization pathway seems to vary according to the NPs and some can be captured and excreted in pseudofaeces (mixture of mucus and undigested particles) before arriving to the stomach, resulting in lower tissue accumulation and higher depuration [177].

The existing in vivo studies utilizing sea urchins mainly focus on the immune status of the animal after exposure to NPs via the injection of the NP suspension into the mouth or directly into the coelomic cavity (through the soft peristomal membrane surrounding the mouth). Consequently, NPs injected orally partially cross the intestinal epithelium, invade the coelomic fluid and are then engulfed by phagocytes, while the remaining particles pass the digestive system and can be excreted [178]. On the other hand, NPs injected into the coelomic cavity can directly interact and be recognized by phagocytes [85,179].

In general, the in vivo passive exposure experiments consider more realistic exposure pathways (feeding, breathing) of NPs. However, for more simplistic set up and to be sure that NPs encounter immune cells, NP suspensions can be also injected into the animal. Results obtained from in vivo tests can provide a good proxy of interactions of immune cells in situ and thereby in vivo tests are crucial to resolve the issue of whether NPs pose an immune threat to living organisms. These models have been shown to be easy to maintain in the laboratory, and exposure experiments allow for the effects of NPs to be studied at different levels of the organism. For future experiments, mesocosms will help to mimic environmental scenarios before further studies in the environment.

#### *3.4. Innate Immune Parameters of Interest*

As highlighted in Figure 7, a variety of endpoints can be used to compare immune system–nanoparticle interactions between different models. This includes functional responses, which comprise the biochemical assessment of cellular and humoral responses, and molecular responses that aim to evaluate changes in the expression of immune-related genes. Because there are many methods available to quantify functional responses, here

we provide a comparative overview of these methods to show which are most appropriate for the purpose of NP testing and cross-species comparisons (see Table 2).

**Figure 7.** Summary of the different endpoints measured in immune cells after exposure to NPs.

**Table 2.** Overview of studies demonstrating the use of cellular and humoral parameters to characterize the immune responses of organisms.



**Table 2.** *Cont.*

#### 3.4.1. Whole Cell Response

Parameters looking at the whole immune cell concerns the immune cells viability, the membrane integrity, all the different types of interactions they can have with NPs, and their potential changes in morphology (Figure 7, first point). The first important cellular responses that can be used to investigate nanoparticle—immune system interactions is (immune) cell viability. Although cell death and apoptosis are part of a normal immune response, studies have shown that NPs are able to cause excessive mortality in immune cells with possible adverse effects on immunocompetence. There are several methods available that can measure cell viability (Table 2). A common method used in human cell lines is the measurement of the release of LDH or ATP through biochemical assays [153,182]. An alternative is the staining of living or dead cells using fluorescent probes (e.g., fluorescein diacetate–FDA or propidium iodide–PI) for observation using flow

cytometry fluorescent microscopy. In some models, cell viability may also be studied by measuring metabolic activity through cell-permeable fluorescent reduction such as CTB or the colorimetric MTT. In some models the use of counter stain dye such as trypan blue or nigrosine [141,192] or the use of DNA-binding florescent dyes are alternative methods for assessing cell viability [150]. There are several methods available to measure the number of cells that are in the process of dying (apoptosis). Apoptosis and preapoptosis evaluation methods, which are available for several models, can be used as early markers for cell viability through the use of specific fluorescent dyes (e.g., annexin V binding, apostain, tetramethylrhodamine, ethylester perchlorate-TMRE or DAPI labelling) [139,197]. Cell viability is probably the best described immune parameter in most species (Table 2); therefore, this parameter is one of most relevant to assess in cross-species comparison. In addition to measuring the overall immune cell viability, quantifying changes in the ratios of different subpopulations of immune cells (e.g., total hemocyte counts-THC) can often give a more detailed view of the impact of NPs on immune cell viability [141,179,199,200]. In human cells, fluorescence-activated cell sorting (FACS) is a common method in which fluorescent antibody-tags can be used to determine a large range of parameters but also to discriminate sub-cell populations [153].

The subcellular effects of NP exposure can be identified via assessment of the integrity of organelles, membranes, and other cellular compartments. Lysosomal functional integrity is an evolutionarily conserved marker of stress (including NPs) and of an individuals' health status, and is commonly evaluated by measuring neural red retention or uptake [185,189,248]. Other approaches that can be used to assess the effects of NPs on organelles include methods measuring trans-golgi apparatus integrity and internal membrane polarization [178,179,200].

Another crucial step in the characterization of NP–immune system interactions is assessing whether immune cells are able to internalize NPs. The internalization of NPs has been observed for different types of NPs across the selected models and was recently reviewed in [36]. There are several techniques available to detect the internalization of NPs. These include transmission electron microscopy (TEM) which can image internalized particles and scanning electron microscopy (SEM) which helps to visualize membranebound particles and can give a direct image of the particles and cells following contact. They can also provide details on how the interaction occurs as well as the state of the NPs (e.g., agglomeration, aggregation, precipitation). These techniques when coupled to an EDX system (energy dispersive X-ray) can be used to perform chemical characterization of the NPs' surfaces. TEM and SEM are descriptive techniques that can provide valuable information but makes quantification difficult between models. Some research has reported the use of fluorescently labelled particles to help to measure particle uptake, although the use of such labelled particles requires additional controls to rule out any effects linked to the leakage of fluorescent dyes [249]. In general, NP internalization in human cells has been well reported but for invertebrate, similar methods often require adjustments to be made (as for example, the salt or osmotic concentrations during fixation) [90,141,153,205,206,209]. For plants, TEM can be used to verify the entry of NPs into the cells [203]. In addition, these kinds of techniques can reveal the change in cell morphologies and subcellular structures (e.g., vacuoles, phagosomes, endosomes) upon NP exposure and give hints regarding the general activation or damages that the cell has undergone [203,207].

#### 3.4.2. Phagocytic Activity

While an immune response towards NPs could be part of normal immune functioning, overstimulation of the immune system resulting in damage or suppression leading to a compromised immune functioning may pose a threat to the organism. Such suppression of immune functioning caused by NPs could be studied via the assessment of the immune cells capacity to phagocytose and the consequent changes on index and rates (Figure 7, second point, Table 2). In earthworms, mussels, and sea urchins, phagocytosis can be evaluated by using fluorescence beads or yeast (using neutral red stained zymosan) [139,145,150,181,216,250].

#### 3.4.3. Cytotoxic Factors

Upon contact with NPs, cells can be activated and produce cytotoxic factors inside the cells in order to help to remove internalized foreign particles (Figure 7, third point). Among them, the oxidative burst, which involves the production of several radicals from oxygen (ROS) and nitrogen (RNS) derivatives. To quantify ROS, several methods, including the use of fluorescent probes (e.g., DCF or calcein), UV-vis spectroscopy (e.g., cytochrome C reduction) or histochemical staining, can be used and many of which have been adjusted for use across the model organisms (Table 2). Moreover, lipid peroxidation can be measured as a proxy for the damages caused by oxidative stress to the membranes, even if it is more frequently analyzed in tissues than in individual cells [138,139]. As for the quantification of RNS and more commonly nitric oxide (NO), in isopods, mussels, and earthworms, NO levels in the hemolymph can be measured spectrophotometrically from hemolymph samples using Griess reagent [141,234,250].

Lysozyme is an evolutionary conserved enzyme that catalyzes the hydrolysis of peptidoglycan and plays a role in the innate immunity of many organisms including earthworms, mussels, sea urchins, and plants [67,234,251,252]. The quantification of the release of lysozyme into the extracellular medium is based on the lysis activity of *Micrococcus lysodeikticus* which can be determined spectrophotometrically. Fluorescent probes can be also used to monitor the evolution of the lysosomal compartment and acidification in the cell upon exposure to NPs [200].

#### 3.4.4. Humoral Factors

Humoral immune responses play a crucial role in immunity by facilitating communication between immune cells and directing the extracellular destruction of foreign objects. Upon activation of the immune cells, some factors can be released into the extracellular medium (Figure 7, fourth point, Table 2). In mammals, cytokine production is a key driver of cellular immune responses [238]. Analyzing the extra- and intracellular concentration of cytokines secreted by (human) immune cells is a well-established method to test the effects and safety of NPs [253]. Many techniques have been developed to detect single or multiple cytokines and factors secreted by cells. These include classic methods such as western blot, ELISA, and bio-chemiluminescence assays, and many commercial possibilities for multiplex assays including legendplex or Ella multiplex technology [153,209,254]. Interestingly, in sea urchins, cytokine IL-6 can be detected in immune cells and secretome using western blot analysis after exposure to NPs [151].

Lysozyme and radicals can also be released by immune cells in the extracellular medium, to directly destruct foreign particles in close proximity to the cell. In addition, there are species specific released factors such as the hemolytic protein lysesin found in the coelomic fluid of earthworms.

Lastly, an important immunological parameter often analyzed in invertebrate models is the measurement of phenoloxidase (PO) activity. This enzyme, produced via the pro-PO cascade, is involved in the production of melanin [43,141]. Upon detecting a melanized pathogen or object, immune cells quickly encapsulate the material resulting in the elimination of the threat. PO activity can be assessed by monitoring the formation of a reddish-brown pigments in the hemolymph from an individual organism using spectrophotometry [44,173,240]. In bivalves, the presence of PO has been reported but its basal levels, the variation across species and especially its response to NP exposure remain poorly understood [83,255]. In the sea urchin *Strongylocentrotus nudus* coelomocytes, three proteins with PO-like activities have been identified using electrophoretic methods [241].

#### 3.4.5. Molecular Response

Increasingly, humoral responses can be measured through genetic or omics approaches (e.g., quantitative PCR or full-transcriptome sequencing) (Table 2). A main advantage of using these approaches over biochemical ones is their high-throughput potential and increased specificity. Furthermore, genetic or omics approaches allow for the assessment of entire immunological pathways instead of focusing on specific biochemical endpoints. However, major limitations of these methods are that they require species specific primers and the availability of transcriptomes, which are currently lacking for many invertebrate species. Moreover, gene expression is highly regulated and time-dependent, so careful consideration must be given to the experimental model in terms of stimulation/exposure time, and cell collection technique.

Genes involved in different immune-related functions, such as oxidative stress response, humoral factors (e.g., AMPs), and receptor proteins, are available for some organisms and the main immune related genes known to be activated upon NP exposure are reported in Table 2.

A whole genome transcript is under development and will be available for *P. scaber*. Using this and genes previously annotated in other more commonly used crustacean species, primers specific for *P. scaber* immune-related genes can be designed (Hernadi, Mayall personal communication). Moreover, gene expression offers alternative possibilities to study several proteins involved in the immune response that biochemical tests that evaluate their activity or functionality are not feasible, such as the effects of NPs on AMP modulation. In plants, microarray-based studies are good tools to monitor the expression of candidate genes involved in the plant defense responses after interaction with NPs [204]. Additionally, an important future issue for environmental molecular biology is to establish whether an up- or downregulation of a certain gene correlates to a modification in the levels of related proteins [256]. The study of transcriptomic changes in cells, tissues, or full invertebrate organisms after exposure to NPs is now emerging but is still in its early phase. Transcriptome analysis can highlight pathways being activated but it should also be accompanied by the study of functional parameters for a fuller, deeper understanding [228,257,258]. In addition, changes in protein repertoires (proteomics approach) have shown interesting outcomes; however, studying of the combined immune response to NP exposure remains in its infancy [259–261].

#### **4. Proposal for Future Cross-Species Evaluations and Conclusions**

During the PANDORA project [35], several studies were conducted on the innate immune response of different models exposed to a large range of different NPs. Based on the outcomes of these studies, several conclusions can be made which may help to guide future (comparative) studies on nanoparticle–immune system interactions (Figure 8).

Because chemical conditions of mediums strongly affect the form and state of NPs and thereby the behavior of NPs, it is crucial to characterize the physico-chemical properties of NPs in the exposure medium as well as in their pristine form (e.g., after production). As the behavior and the interaction of NPs with immune systems are also time-dependent, experimental design will need to critically consider exposure duration as well. In vitro models can be considered as the prime focus for studying nanoparticle–immune system interactions. However, in vitro models are not available for all immune model species (e.g., isopods, plants), limiting comparative studies based on in vitro testing. In vivo experiments are crucial to study nanoparticle–immune system interactions under more realistic conditions. Studies using invertebrates, which are well-established and can be conducted on a routine basis, may serve as good alternatives to in vivo mammalian testing models such as mice or rats.

Here, exposure route and exposure concentration will need to be critically considered as these factors are likely to significantly affect immune cells and their interaction with the NPs.

**Figure 8.** Proposal template for translatable NP experiments across the models of interest (plants, terrestrial and marine invertebrates, and human cells).

> In this review, we provided an overview of the methods used to characterize nanoparticle immune responses in various organisms across the tree of life (Table 2). Among cellular parameters, it appears that methods to assess cell viability (including assessments of subpopulations) and NPs internalization by immune cells are well described in most organisms. Phagocytic activity is a crucial parameter to be evaluated for immune cells, however, some models lack the methods to study this parameter (e.g., isopods) or do not rely on this type of response (e.g., plants). Moreover, as methods for microscopy are universally available, measurements of the change in morphology and external interactions of NPs can be studied in most immune models.

> Due to a lack of general knowledge on the composition and functioning of humoral immunity in most organisms other than mammal/human models, it remains difficult to identify the most relevant humoral parameter for cross-species evaluations. The exception being oxidative (and nitrosative) stress, for which methods are well described in most species.

> Further work is needed to identify the interorganism comparability of otherwise species-specific markers, especially in invertebrates. In order to fully characterize NP-immune responses across species from the tree of life, there is a need for the iden

tification of markers indicative for both pro- and anti-inflammatory responses, as are currently already available for human models (e.g., [253]). The development of such markers will require fundamental research on the innate immune systems of organisms other than human models. Thorough investigations in species from across the tree of life will help to understand how NPs interact with the innate immune system under different conditions and environments which may guide the future development of NPs that are immunologically safer-by-design.

**Author Contributions:** Conceptualization, B.J.S., C.M. and M.A.; writing—original draft preparation, C.M., B.J.S., A.A., F.B., E.F., S.H., S.M., N.I.N.P., A.P., E.S., M.A.; writing—review and editing, B.J.S., C.M., E.S., M.A.; supervision, M.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** All authors were supported by the EU H2020 project PANDORA, grant number 671881.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** The authors warmly acknowledge the project coordinator Diana Boraschi for her strong support during the whole project period, and all the PIs (L. Canesi, D. Drobne, A. Duschl, M.A. Ewart, J. Horejs-Hoeck, P. Italiani, B. Kemmerling, P. Kille, A. Pinsino, P. Procházková, V.F. Puntes, D.J. Spurgeon, C. Svendsen, C.J. Wilde) for the offered opportunity. The authors wish also to acknowledge Paola Cesaroni for her immense organizational help. Special thanks to our Project Officer Giuliana Donini for making PANDORA a successful story.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Nanomaterials* Editorial Office E-mail: nanomaterials@mdpi.com www.mdpi.com/journal/nanomaterials

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com

ISBN 978-3-0365-3948-5