A representative “long” and “thin” MWCNT was selected for testing in the Division of Translational Toxicology (DTT), NIEHS Immunotoxicology studies. The 1020 Long Multiwalled Carbon Nanotube (L-MWNT-1020) from Sun Innovations (Fremont, CA) was selected based on commercial availability of large quantities, high purity (98%), and the low amount of residual metal (nickel) catalyst (0.52% by weight). Identity and purity analyses were conducted by various analytical chemistry techniques (NTP, 2019) and the L-MWNT-1020 was stored in the original shipping containers at room temperature (approximately 25°C). Reanalysis of the lot 10031301M used for these studies showed an average length and width of 2,400 nm and 16 nm, respectively, which was in accordance with the characterization by the manufacturer as 10–30 µm length and 10–20 nm diameter. The L-MWNT-1020 was composed almost entirely of carbon and was stable and chemically unreactive; therefore, the potential pulmonary toxicity and hazard of L-MWNT-1020 should be primarily a function of physical dimensions, exposure concentration, and biopersistence.

All animal procedures were approved by the institutional animal care and use committee (IACUC) at the appropriate institution, Battelle Memorial Institute (Headquarters, Columbus, OH) for the inhalation exposure phase, and Burleson Research Technologies, Inc. for the immunotoxicity assessments. Exposures were conducted by Battelle Memorial Institute using whole body inhalation systems as previously described in detail (Ihrie et al., 2019; NTP, 2019). Briefly, the aerosol generation system consisted of a linear feed dust metering device to m L-MWNT-1020 from a reservoir into an air stream. A particle attrition chamber was positioned immediately downstream of the metering device exhaust tube to reduce the particle size of the test material. A single jet disperser assisted in further dispersion and particle size reduction and a cyclone separator removed the larger particles from the distribution system. Generated aerosol was conveyed from an exposure control suite to the exposure room. The inhalation exposure chamber (Lab Products, Inc., Seaford, DE) was designed so uniform aerosol concentrations could be maintained throughout the chamber with a total active mixing volume for each chamber of 1.7 m³. The concentrations of L-MWNT-1020 in the exposure chambers and room air were monitored using three real-time aerosol monitors. Aerosol particle size distribution and uniformity of aerosol concentration in the inhalation exposure chambers without animals present were evaluated before the studies began (NTP, 2019). Female mice and rats were used for these studies as they tend to be more sensitive to immunotoxicity and show greater dynamic range in immune responses than males (Klein and Flanagan, 2016). Female B6C3F1/N mice (Taconic Biosciences, Hudson, NY) were exposed for 30 or 90 days to L-MWNT-1020 at concentrations of 0, 0.06, 0.2 or 0.6 mg/m³ while cohorts of female Hsd:Harlan Sprague Dawley SD^® rats were exposed for 30 days at the same concentrations 6 h/day for 5 days/week, as shown in Figure 1. Following the prescribed exposure period, animals were transferred to Burleson Research Technologies, Inc. (BRT; Morrisville, NC) to provide immunizations or infections and subsequent immune function testing ex vivo. A positive control group was also included in which mice received 50 mg/kg bodyweight of cyclophosphamide (CPS) by intraperitoneal injection, once per day for 4–14 days (cohort dependent), prior to scheduled euthanasia and rats received 15 mg/kg bodyweight, once per day for 4–14 days (cohort dependent), prior to assessment of immune function. All animals were euthanized by CO₂ inhalation using 100% CO₂ introduced at 3.65 LPM into a 7.3 L chamber to displace 50% of the atmosphere per minute until breathing ceased and no pedal reflex was observed. Severing the diaphragm was used for confirmation of death.

Animals were single housed in individually ventilated cages and acclimated to the BRT vivarium for at least 7 days. Municipal tap water (Town of Cary, NC) and NTP-2000 pelleted diet were provided ad libitum. All animals were kept in facilities with a 12-h light/dark cycle at 19.3°C–25.0°C and with a 30%–64% relative humidity. Following acclimation, animals were randomized by bodyweight (±20% of mean bodyweight) to six independent cohorts (Table 1) for each study (30-day mouse, 90-day mouse, and 30-day rat studies). Each cohort for the mouse studies consisted of 8 animals/treatment group unless otherwise stated. Each cohort for the rat study consisted of 12 rats/treatment group with the exception of the CPS groups which consisted of 10 rats/group.

Blood was collected by the retro-orbital (RO) route in anesthetized animals into EDTA blood collection tubes for analysis (complete hematological profiles, white blood cell differentials and reticulocyte counts). Due to unsuccessful RO blood collection, several animals had blood collected via cardiac puncture. The blood was analyzed the day of collection on an Advia 120 hematology analyzer using associated V.6.3.2-MS software (Siemens Medical Solutions United States, Inc., Malvern, PA).

Following completion of blood collection, animals were euthanized with CO₂ and selected tissues for histopathology were collected. Gross observations were documented and the spleen, thymus, liver, right kidney with adrenals, and lungs were weighed. The spleen, thymus, mesenteric and popliteal lymph nodes, and liver (median, caudate, and right lobes) were preserved in 10% neutral buffered formalin (NBF). Other organs and tissues showing gross lesions were preserved in 10% NBF. Thymus, spleen, lymph nodes, bronchoalveolar lymphoid tissue (BALT), and bone marrow from the femur were examined using an enhanced histopathology method (Elmore, 2012), while the lung, liver, kidney (right), and adrenal gland (right) were evaluated using routine histopathology. Bone marrow cells from the right femur were removed and collected for total cell counts and cytospin preparations.

BALF was collected, and mononuclear cells (50,000 cells) were mixed with opsonized staphylococcus aureus bioparticles (100 μg) labeled with pH sensitive pHrodo™ Red fluorochrome (Invitrogen; Eugene, OR) that fluoresces red within phagosomes due to the acidic pH. Cytochalasin D (15 μM; Sigma, St. Louis, MO), an in vitro inhibitor of phagocytosis, was added to selected samples for 45 min prior to the addition of bioparticles. Cells and bioparticles were incubated for 2 h (±5 min) and then cooled to 2°C–8°C to stop phagocytosis. Phagocytosis was quantified as the increase in cell-associated red fluorescence (peak emission at 587 nm) using an Accuri C6^® flow cytometer using CFlow Plus^® software v1.0.264.15. Data were expressed as the percentage of cells that engulfed bacteria and the mean fluorescent intensity of cells that engulfed bacteria (measure of number of bacteria engulfed per cell).

Humoral mediated immunity was assessed using two model T-dependent antigens, sheep red blood cells (SRBC, Colorado Serum Company, Denver, CO) and keyhole limpet hemocyanin (KLH; GMP-grade whole subunit, Stellar Biotechnologies, Inc., Los Angeles, CA). Four days following intravenous immunization with SRBC (7.5 × 10⁷/mouse and 1 × 10⁸/rat), the T-dependent antibody response (TDAR) was assessed by measuring the number of antibody forming cells (AFC) in the spleen using the hemolytic plaque assay, the number of IgM antibody producing B cells using ELISpot, and the serum titers of circulating IgM antibody to SRBC. A separate group received intraperitoneal immunization with KLH 14 days prior to euthanasia and serum IgM and IgG antibody titers determined 5 (IgM) and 14 (IgG) days following immunization. We have described these methods in detail previously (Watson et al., 2021).

T-cell proliferation was examined in single cell suspensions from the spleen in response to ex vivo treatment with species specific monoclonal anti-CD3 antibodies (BD Bioscience, Franklin Lake, NJ), as previously described (Watson et al., 2021). Briefly, microtiter plates were coated overnight with anti-mouse CD3 antibody and then washed. Splenocytes in RPMI complete medium were added to the appropriate wells and incubated at 37°C and 5% CO₂ for up to 96 h. A non-radioactive assay system was used to determine T-cell proliferation (CyQuant Direct Cell Proliferation Assay^®, Molecular Probes, Eugene, OR) according to the manufacturer’s instructions. Fluorescence signal is directly proportional to DNA content and cell number. Data were collected using a SpectraMax M2e spectrofluorometer running Softmax Pro^® (Molecular Devices, San Jose, CA) software v5.0.1.

To stimulate an in vivo cell-mediated immune response, mice and rats from Cohort 5 were infected with influenza virus [∼4 × 10⁴ plaque forming units (pfu)/mouse or ∼2 × 10⁵ pfu/rat] via intranasal instillation (Burleson et al., 2018; Watson et al., 2021) 8 days prior to scheduled termination. The ex vivo CTL assay was performed using lung effector cells isolated from the influenza-exposed animals 8 days following infection as previously described (Watson et al., 2021). Briefly, on the day of the assay either EL-4 (mice) or UMR-106 (rat) target cells (ATCC, Manassas, VA) were infected with influenza virus (multiplicity of 10) in vitro and then labeled with Chromium-51 (⁵¹Cr at 100 μCi per 1 × 10⁶ target cells for 90 min). Lung effector cells (isolated from bronchoalveolar lavage fluid from mice or lung tissue digests from rats) were separated from red blood cells and adherent cells and combined with labeled target cells in U-bottomed microtiter plates at the appropriate effector-to-target ratios (20:1, 10:1, and 5:1 for mice or 50:1, 25:1, and 12.5:1 for rat). Plates were briefly centrifuged to facilitate effector-to-target cell contact and incubated at 37°C/5%CO2 for 6 h. Culture supernatants were harvested and release of ⁵¹Cr was determined using a Cobra II Auto-Gamma counter (Packard Inc., Pamsey, MN). Specific target cell lysis is a direct measure of influenza-specific CTL killing activity.

To stimulate in vivo recruitment and activation of NK cells in the lung, mice or rats from cohort 6 were infected with influenza virus [∼4 × 10⁴ plaque forming units (pfu)/mouse or ∼2 × 10⁵ pfu/rat] via intranasal instillation (Watson et al., 2021) 2 days prior to scheduled termination for rats and 3–4 days prior to scheduled termination for mice. Pulmonary NK cell killing activity was evaluated using YAC-1 tumor target cells as previously described (Watson et al., 2021). Briefly, lung effector cells (100 μL) were added to wells of round-bottom microtiter plates containing 100 μL of YAC-1 target cells (1 × 10⁵ cells/mL labeled with ⁵¹Cr at 100 μCi per 1 × 10⁶ target cells for 90 min) at effector to target ratios of 20:1, 10:1, and 5:1 for mice or 50:1, 25:1, and 12.5:1 for rats. The plates were centrifuged and then incubated at 37°C/5%CO₂ for 4 h. Culture supernatants were harvested and release of ⁵¹Cr was determined using a Cobra II Auto-Gamma counter (Packard).

Splenic immune cell populations were determined using flow cytometry and are presented as absolute cell numbers and relative percentages of CD45⁺ lymphocytes or total myeloid cells, as appropriate. Red blood cells (RBC) present in single cell suspensions from the spleen were removed by lysis prior to Fc receptor blocking and antibody staining. Detailed procedures including antibodies, staining, and gating strategies were previously published (Watson et al., 2021). Antibody cocktails for discrimination of immune cell populations were the following.

• Rat: (1) CD45, CD3, CD45RA, and CD161a; (2) CD45, CD3, CD4, CD8a; and (3) CD11b/c, CD103, RP-1, and CD161a.

Details of the antibodies used for these studies are provided in the CEBS Supplemental Antibody Table (10.22427/NTP-DATA-500-005-001-000-1). Cell populations examined in the spleen included total lymphocytes, total T cells, CD4⁺ T cells, CD8⁺ T cells, B cells, NK cells, monocytes/macrophages, eosinophils, and neutrophils. In addition, T:B cell and CD4+:CD8+ T cell ratios were determined. Following staining, samples were stored at 2–8°C, protected from light, until analyzed on an Accuri C6 flow cytometer using CFlow Plus v 1.0.264.15 (BD Biosciences). In all cases, a minimum of 20,000 events/sample was acquired.

Lymphocyte gating was performed on CD45⁺ populations. The following lymphocyte subsets were identified; T-cells (CD3⁺CD45RA⁻), B-cells (CD3⁻CD45RA⁺), NK cells (CD3⁻CD161a⁺), T-helper cells (CD3⁺CD4⁺), and T-cytotoxic cells (CD3⁺CD8⁺). Myeloid cells were gated based on being positive for CD11b with low-to-mid intensity staining for CD11c (mouse) or CD103 (rats). Myeloid populations were differentiated from NK cells based on lack of CD335(NKp46) expression, (mouse) or lack of CD161a (rats). Further differentiation was based on expression of Ly6G (mouse) or RP-1 (rat) with positive cells being neutrophils and negative cells differentiated using SSC into monocytes/macrophages with low granularity and eosinophils with high granularity.

Data were collected into Provantis v9.3.2.1 (Instem, Philadelphia, PA) and calculation of endpoints was performed within this validated electronic data collection and management system. Results are presented as mean ± SEM. Extreme values were identified by the outlier test of Dixon and Massey when sample size was <20 (Dixon and Massey, 1957) and by Tukey’s outer fences method (Tukey, 1977) when sample size was ≥20. All flagged outliers were examined by DTT, NIEHS personnel, and implausible values were eliminated from the final analyses. Jonckheere’s test was used to test for dose-related trends (Jonckheere, 1954). The positive control group was excluded from the trend tests. Bodyweight and organ weight data, which exhibit a normal distribution, were analyzed using Jonckheere’s trend test (Jonckheere, 1954) and Williams’ or Dunnett’s (pairwise) test depending on detection of a trend at a 0.01 significance level (Dunnett, 1955; Williams, 1971; Williams, 1972; Williams, 1986). Data for other endpoints were analyzed using a non-parametric multiple comparison procedure. Shirley’s test was used if a significant trend was observed (Shirley, 1977) and Dunn’s test was used if the trend was not significant (Dunn, 1964). Positive control group data were compared to the vehicle control group using the t-test for bodyweight and organ weight and Wilcoxon rank-sum test for all other endpoints. Data that were different from control at p ≤ 0.05 were considered statistically significant.

In all three experimental studies occasional minor clinical abnormalities were observed, such as alopecia in mice, but were not considered related to exposure as similar incidences were observed between control and treated groups (CEBS Summary Table I05). The exposures to L-MWNT-1020 were performed by the inhalation contractor followed by shipment of the animals to BRT for assessment of immune function. L-MWNT-1020 inhalation had minimal effects on in-life bodyweights in mice and rats (CEBS Summary Tables I04 and I04G). Exposures were not continued at BRT and there were no effects on terminal bodyweights in any of the cohorts. At necropsy, the thymus, spleen, liver, right kidney with adrenals, and lungs were removed, examined, and weighed. There were no significant gross pathology (CEBS Summary Table PA46) or organ weight changes (CEBS Summary Table PA06) in any of the organs collected when compared to the vehicle control group. Histological examination revealed low numbers of pulmonary macrophages that contained black pigment, consisting of agglomerated L-MWCT-1020, in the majority of rats treated for 30 days with ≥0.2 mg/m³ and in all mice from the 0.6 mg/m³ treatment group following 90 days of exposure, but not in mice treated for 30 days. There were no other histopathological changes that could be attributed to inhalation of L-MWNT-1020 in any lymphoid or non-lymphoid tissue examined in either mice or rats (CEBS Summary Table PA02, PA03, PA08, and PA10).

L-MWNT-1020 inhalation resulted in small, but statistically significant changes in hematological values and white blood cell (WBC) differentials. There was a significant decreasing trend with increasing exposure concentration in platelet counts in mice treated for 90 days and a significant decrease in mice treated with 0.6 mg/m³ L-MWNT-1020 compared to the vehicle control group. A similar negative trend in platelets was observed in rats treated for 30 days, however, no treatment groups were different from the vehicle group (CEBS Summary Table M04). Examination of WBC differentials (Table 2, CEBS Summary Table M03) indicated that there were significant increasing trends in absolute and relative neutrophils and eosinophils in rats exposed to L-MWNT-1020 for 30 days. There was also a decreasing trend in the relative lymphocyte counts in exposed rats. These trends were accompanied by significantly increased absolute neutrophil counts (0.6 mg/m³), relative neutrophil counts (≥0.2 mg/m³), and absolute and relative eosinophil counts (≥0.2 mg/m³), as well as significantly decreased relative lymphocyte counts (≥0.2 mg/m³) in rats exposed to L-MWNT-1020 for 30 days. Mice exposed to L-MWNT-1020 for 30 days appeared less sensitive and only had significantly decreased numbers of monocytes and neutrophils in the 0.06 mg/m³ treatment group. Extending the exposure period in mice to 90 days resulted in a significantly increasing trend in relative neutrophil counts which was significantly increased in the 0.6 mg/m³ group relative to the vehicle group which is consistent with the findings in rats. Mice exposed to L-MWNT-1020 for 90 days also showed significantly decreasing trends in absolute lymphocytes, basophils, and large unstained cells but only large unstained cells in mice exposed to 0.6 mg/m³ were significantly decreased relative to the vehicle control group (CEBS Summary Table M03). No other changes were observed in hematology parameters in mice or rats treated with L-MWNT-1020. The positive control, CPS, produced a hemogram consistent with severe leukocytopenia as expected.

The impact of L-MWNT-1020 exposure on the number and ability of bronchoalveolar lavage fluid (BALF) mononuclear cells to phagocytize opsonized Staphylococcus aureus bioparticles was determined (Figure 2; CEBS Summary Table M14). In mice exposed for 30 days, there was a significant increase in the number of pulmonary mononuclear cells present in the BALF in the 0.6 mg/m³ exposure group (Figure 2A) and in the percentage of mononuclear phagocytic cells actively engulfing opsonized staphylococcus aureus bioparticles in both the 0.2 and 0.6 mg/m³ exposure groups (Figure 2B). The relative number of opsonized Staphylococcus aureus bioparticles phagocytized per cell, as measured by mean fluorescent intensity (MFI), was not affected by exposure in mice (Figure 2C). Neither pulmonary mononuclear cell number nor phagocytosis were affected in rats exposed by inhalation for 30 days to L-MWNT-1020 (Figures 2D–F). There was insufficient uptake of the fluorescent bioparticles by the pulmonary phagocytic cells for all animals and assessment of phagocytic function was not possible in the 90-Day mouse study (data are not presented).

L-MWNT-1020 exposure in rats significantly decreased the anti-SRBC AFC response (CEBS Summary Table M07) at the 0.2 mg/m³ dose level when expressed as AFC/10⁶ spleen cells (Figure 3E). A decrease was also observed when expressed as AFC/spleen but did not reach statistical significance (Figure 3F). While the responses were also decreased at the 0.06 and 0.6 mg/m³ dose levels in rats, statistical significance was not achieved. Similar decreases were observed in the AFC responses in mice, but these changes did not reach statistical significance (Figures 3A–D). None of these groups showed significant dose response trends. Serum levels of anti-SRBC IgM antibodies (CEBS Summary Table M08) and anti-SRBC IgM producing B-cells (CEBS Summary Table M19) were negligibly impacted by inhalation of L-MWNT-1020 in mice and rats. In addition, antibody responses to KLH were not affected by L-MWNT-1020 inhalation in rats or mice (CEBS Summary Table M09). Humoral immune responses to SRBC and KLH were markedly decreased in all CPS treated animals as expected.

Cell mediated immune function was assessed by quantitating splenic T cell proliferation in response to stimulation with anti-CD3 monoclonal antibodies and by determining the ability of pulmonary T cells from influenza-infected animals to lyse influenza-infected target cells in a cytotoxic T lymphocyte (CTL) assay. There was a statistically significant increase in anti-CD3 antibody stimulated T cell proliferation at the lowest (0.06 mg/m³) dose of L-MWNT-1020 in mice exposed for 30 days but no effect was observed after 90 days in mice. A significant decrease in proliferation occurred in rats exposed to L-MWNT-1020 for 30 days, but only in the 0.2 mg/m³ treatment group (CEBS Summary Table M11). These changes in T-cell proliferation were minor and not dose-related. CTL activity was unaffected following exposure to L-MWNT-1020 in either mice or rats (CEBS Summary Table M12). CPS treatment markedly decreased the CTL response as expected.

Natural killer (NK) cell activity was evaluated by examining the ability of pulmonary effector cells isolated 2 days following influenza infection to lyse YAC-1 tumor target cells. There was a significant decreasing trend in NK killing activity in mice exposed to L-MWNT-1020 for 90 days but only at the 5:1 E:T ratio. This trend was not observed in mice or rats exposed for 30 days. There were no other exposure-related effects on NK activity following inhalation of L-MWNT-1020 in either rats or mice (CEBS Summary Table M15).

The absolute number and relative percentages of immune cell populations in the spleen are presented in Tables 3, 4, respectively. Mice exposed to L-MWNT-1020 for 90 days showed a significant decreasing trend in relative percentage and absolute number of neutrophils with both measures being significantly decreased in the 0.6 mg/m³ treatment group. Changes in other cell populations in the spleen were mostly negligible in mice treated for 30 or 90 days. In rats exposed to 0.06 and 0.2 mg/m³ L-MWNT-1020 for 30 days there was a significant increase in the absolute number of spleen cells that corresponded with a significant increase in total lymphocytes in the same groups. The impact of L-MWNT-1020 on immune cell population numbers was on the T cell compartment as evidenced by significant increases in total T cells, CD4⁺ T cells, CD8⁺ T cells, and the total T cell:B cell ratios in rats treated with 0.06 and 0.2 mg/m³. In addition, there was a positive trend in NK cell numbers with all exposure levels being significantly elevated. CPS treatment was consistent with historical controls resulting in a marked reduction in the number of almost all cell phenotypes with B cells being the most affected and NK cells and monocytic cells the least.