Registration Dossier

Toxicological information

Repeated dose toxicity: inhalation

Currently viewing:

Administrative data

sub-chronic toxicity: inhalation
other: subchronic with recovery (up to 6 months)
Type of information:
experimental study
Adequacy of study:
key study
1 (reliable without restriction)
Rationale for reliability incl. deficiencies:
other: GLP guideline study

Data source

Referenceopen allclose all

Reference Type:
study report
Report Date:
Reference Type:
Subchronic 13-week Inhalation Exposure of rats to Multiwalled Carbon Nanotubes: Toxic effects are determined by density of agglomerate structures, not fibrillar structures
Pauluhn J
Bibliographic source:
Toxicological Sciences 113 (1), 226-242

Materials and methods

Test guidelineopen allclose all
according to
OECD Guideline 413 (Subchronic Inhalation Toxicity: 90-Day Study)
according to
EU Method B.29 (Sub-Chronic Inhalation Toxicity:90-Day Study)
according to
EPA OPPTS 870.3465 (90-Day Inhalation Toxicity)
GLP compliance:
yes (incl. certificate)

Test material


Test animals

Details on test animals and environmental conditions:
- Source: Harlan-Winkelmann (Borchen, Germany)
- Age at study initiation: approx. 2 months
- Weight at study initiation: males: 226-231 g; females: 178-183 g
- Temperature (°C): 22 +/- 3
- Humidity (%): 40-60
- Air changes (per hr): approx. 10
- Photoperiod (hrs dark / hrs light): 12 / 12

Administration / exposure

Route of administration:
inhalation: dust
Type of inhalation exposure:
nose only
Remarks on MMAD:
MMAD / GSD: Berner Cascade Impactor:

0.1 mg/m³,: not applicable
0.4 mg/m³, MMAD (GSD): 3.05 µm (1.98)
1.5 mg/m³, MMAD (GSD): 2.74 µm (2.11)
6.0 mg/m³, MMAD (GSD): 3.42 µm (2.14)

Laser velocimeter TSI APS 3321:

0.1 mg/m³, MMAD (GSD): 1.67 (0.9)
0.4 mg/m³, MMAD (GSD): 1.91 µm (0.93)
1.5 mg/m³, MMAD (GSD): 1.93 µm (0.93)
6.0 mg/m³, MMAD (GSD): 2.19 µm (0.94)
Details on inhalation exposure:

- System of generating particulates:

The test substance was micronized with a Retsch Centrifugal Ball Mill S100. The test substance was characterized before and after micronization. This comparison demonstrated that the test article was not affected by the micronization process to any significant extent. In brief, the reanalysis of the micronized test article showed that the fraction of particulates below 4 µm was increased as called for by the testing requirements. This was achieved without changing the composition and morphology of the agglomerated structures of the actual test substance. A confirmatory analysis was also made from the aerosolized test article (collection of particulates from the inhalation chamber). Again the composition and morphology of the agglomerated structures of the aerosolized test article were essentially identical with the non-dispersed/nonmicronized test substance.

- Description of analytical method used for verification of test substance stability:

Laser Diffraction
Scanning Electron Microscopy (SEM)
Inductive Coupled Plasma - Optical Emission Spectroscopy (ICP-OES)
X-ray Electron Spectroscopy (XPS)
Surface Area Burnauer Emmett-Teller (BET)
Transmission Electron Microscopy (TEM)

Animals were exposed to the aerosolized test substance in Plexiglas exposure restrainers. Restrainers were chosen that accommodated the animals' size. These restrainers were designed so that the rat's tail remained outside the restrainer, thus restrained-induced hyperthermia can be avoided. This type of exposure principle is comparable with a directed-flow exposure design and is preferable to whole-body exposure on scientific and technical reasons (rapid attainment of steady-state concentrations, no technical problems with regard to test atmosphere inhomogeneities, better capabilities to control all inhalation chamber parameters, easier cleaning of exhaust air, and lower consumption of test substance). Moreover, contamination of the haircoat can largely be avoided and confounding effects as a result of uptake of test substance by non-inhalation routes are minimized. The chambers used are commercially available (TSE, 61348 Bad Homburg) and the performance as weil as their validation has been published.

Dry conditioned air was used to eva po rate the test substance as described above. The test atmosphere was then foreed through openings in the inner concentrie cylinder of the ehamber, direetly towards the rats' breathing zone. This directed-flow arrangement minimizes re-breathing of exhaled test atmosphere. Each inhalation ehamber segment is suitable to aecommodate 20 rats at the perimeter loeation. All air f10ws were monitored and adjusted continuously by means of calibrated and computer eontrolled mass-f1ow-controllers. A digitally controlled ealibration f10w meter was used to monitor the aceuraey of mass-flow-eontroller. As demonstrated in Table 1, the ratio between supply and exhaust air was seleeted so that 90% of the supplied air was extraeted via the exhaust air loeation and, if applieable, via sampling ports. Gas serubbing devices were used for exhaust air
elean-up. During sampling, the exhaust air was reduced in accordance with the sampling flow rate using a computerized Data Acquisition and Control System so that the total exhaust air ftow rate was adjusted on-line and maintained at the speelfled 90%. The slight positive balance between the air volume supplied and extracted ensured that no passive Inftux of air into the exposure chamber occurred (via exposure restrainers or other apertures). The slight ~itive balance provides also adequate dead-space ventilation of the exposure restrainers. The pressure difference between the inner inhalation ehamber and the exposure zone was 0.02 cm H20. The exposure system was accommodated in an adequately ventilated enelosure. Temperature and humidity are measured by the Data Acquisition and Contra! System using eallbrated sensors.

The aluminum inhalation chamber has the following dimensions: inner diameter = 14 cm, outer diameter.= 35 cm (two-chamber system), height = 25 cm (internal volume = about 3.8 L). To be able to perform all measurements required to define exposure in a manner that is similar to the exposure of rats, a 'four segment' chamber were used in all groups. Details of this modular chamber and its validation have been published previously.

The test atmosphere generation conditions provide an adequate number of air exchanges per hour [60 L/min x 60 min/(4 x 3.8 L/chamber) = 237, continuous generation of test atmosphere]. Based on OECD-GD39 the equilibrium concentration (t95) can be calculated as folIows:
t95 (mln) = 3x (chamber volume/chamber airflow)
Under the test conditions used a chamber equilibrium is attained in less than one minute of exposure. At each exposure port a minimal air flow rate of 0.75 L/min was provided. The test atmosphere can by no means be diluted by bias-air-flows.

Compressed air was supplied by Boge compressors and was conditioned (i.e. freed from water, dust, and oil) automatically by a VIA compressed air dryer. Adequate control devices were employed to control supply pressure.

Air flows are monitored and controlled continuously by calibrated mass flow meters (Hastings HFC-C Mass Flow Controllers, Teledyne Hastings-Raydist, Hampton, VA, USA). For analytical sampling TYLAN FC-280 S mass flow controllers are used (TYLAN General, Torrance, California, USA). The proper performance of the mass flow controllers was measured using a digital precision flow-meter calibration device (DryCal DC-2 primary gas flow meter). The calibration of mass flow controllers is performed by computer under actual operating conditions. Voltage specifications exceeding or falling below the specified range are indicated by an alarm/error listing. The Data Acquisition and Control System monitors/controls up to five inhalation chambers simultaneously.

The exhaust air was purified via filter mat/Makropur F-filter, cotton-wool + HEPA filters.

Temperature and humidity measurements are also performed by the computerized Data Acquisition and Control System using FTF11 sensors (ELKA ELEKTRONIK, Lüdenscheid, Germany). The position of the probe was at the exposure location of rats. Measurements were performed in the exhaust air. Temperature and humidity data are integrated for 30-seconds and displayed accordingly. The humidity sensors are calibrated using saturated salt solutions according to Greenspan (1977) and Pauluhn (1994) in a two-point calibration at 33% (MgCI2) and at 75% (NaCI) relative humidity. The calibration of the temperature sensors is also checked at two temperatures using reference thermometers.
Analytical verification of doses or concentrations:
Details on analytical verification of doses or concentrations:

- General remark: A nominal concentration was not calculated since the construction and weight of the dust generator used did not allow for a precise measurement of the powder aerosolized.
- Gravimetric evaluation: The test-substanee concentration was determined by gravimetric analysis (filter: Cellulose-Acetate-Filter, Sartorius, Göttingen, Germany; balance: Mettler AE 100). Chamber samples were taken in the vicinity of the breathing zone. The number of sampIes taken was sufficient to characterize the test atmosphere and was adjusted so as to accommodate the sampling duration and/or the need to confirm specific concentration values. Optimally, samples were collected in hourly intervals. All analytical concentrations reported refer to mg of test substance/m3 air (gravimetrical method).

The integrity stability of the aerosol generation system was monitored using a Mierodust Pro (Casella) digital real-time aerosol photometer. In addition, at 0.1 mglm3 only, a TEOM 1400a (Rupprecht & Patashnick Co., Ine.) system was used. The latter provides a means for real-time, true msss measurement of suspended partieulate matter. This instrument incorporates the patented tapered element oscillating microbalance, a true microweighing technology that provides true mass measurements. This mieroproeessor-based unit easily accommodates an inertial balance that directly measures the msss collected on an exehangeable filter cartridge by monitoring the corresponding frequency ehanges of a tapered element. The sampie flow passes through the filter, where particulate matter collects, and then continues through the hollow tapered element on its way to an active volumetrie flow control system and vacuum pump. Active volumetrie flow control is maintained by mass flow controllers whose set points are constantly adjusted in accordance with the measured ambient temperature and pressure. All sampies were taken continuously from the vicinity of the breathing zone. The results are displayed on the computer screen and printed after cessation of exposure. For data recording and display the system integration time is 30 sec. This chamber monitoring allows for an overall survey of toxicologically relevant technical parameters (inlet and exhaust flows as weil as atmosphere homogeneity, temporal stability, and generation performance). Interruptions in exposure (e.g. resulting from obstruction of nozzles or other technical mishaps) are recorded and, if
applicable, a commensurate interval is added to the exposure duration for compensation.

The sampies for the analysis of the particle-size distribution were also taken in the vicinity of the breathing zone . Cascade impactor analyses were not
feasible at 0.1 mg/m³ as this would had required extensive sampling period. Therefore, a TSI-Laser Velocimeter APS 3321 was used for particle size analyses in this group. Due to the entirely different physical principle of measurement, this analysis was also performed in the remaining exposure groups for comparison. TSI-Laser Ve/ocimeter APS 3321, Including diluter TSI Model 3302 (TSllnc., St Paul, MN, USA). Technical details of this system have been described elsewhere. The APS 3321 was used especially for determinations at low concentrations. Analyses performed at higher concentrations were made to allow
comparisons with cascade impactor analyses.

The parameters characterizing the particle-size distribution were calculated according to the following procedure:
Mass Median Aerodynamic Diameter (MMAD): Construct a 'Cumulative Percent Found - Less Than Stated Particle Size' table, calculate the total mass of test substance collected in the cascade impactar. Start with the test substance collected on the stage that captures the smallest particle-size fraction, and divide this mass of the test substance by the total mass found above. Multiply this quotient by 100 to convert to percent. Enter this percent opposite the effective cut-off diameter of the stage above it in the impactor stack. Repeat this step for each of the remaining stages in ascending order. For each stage, add the percentage of mass found to the percentage of mass of the stages below it. Plot the percentage of mass less than the stated size versus particle size in a probability scale against a log particle-size scale, and draw a straight line best fitting the plotted points. A weighted least square regression analysis may be used to achieve the best fit. Note the particle size at which the line crosses the 50% mark. This is the estimated Mass Median Aerodynamic Diameter (MMAD).

Duration of treatment / exposure:
13 weeks
Frequency of treatment:
6 hours/day, 5 days/week
Doses / concentrationsopen allclose all
Doses / Concentrations:
0.1, 0.4, 1.5, 6.0 mg/m3
other: target concentration
Doses / Concentrations:
0.10, 0.45, 1.62, 3.66 mg/m3
other: gravimetric concentration
No. of animals per sex per dose:
40 males and 10 females per group
Control animals:
yes, sham-exposed
Details on study design:
Based on the rational described in the publication (Pauluhn J, Toxicol. Sciences 113 (1), 226-242 (2010) MWCNT should be approx. 10 times more potent than carbon black due to their higher displacement volume in alveolar macrophages and 0.1, 0.4, 1.5 and 6 mg/m³ were selected as target concentrations. Based on this hypothesis, 0.1 mg/m³ is estimated to be in the nonoverload range and 0.4 mg/m³ is assumed to be at that transitional level at which lung overload may occur. Frank lung overload with partial or inhibited clearance is anticipated to occur at 1.5 and 6 mg/m³.

Recovery groups of male rats were sacrificed 4, 13 and 26 weeks after the end of the exposure period, respectively.


Observations and examinations performed and frequency:
Body weights of all animals were measured before exposure on a twice per week basis. During the exposure-free recovery period the body weights were determined once per week.

Food and water consumption were determined on a per week basis.

The appearance and behavior of each rat was examined carefully at least twice on exposure days (before and after exposure) and once a day on exposure-free days. If considered applicable due to unequivocal signs, in nose-only exposed rats observations were also made during exposure. Following exposure, observations were made and recorded systematically; individual records were maintained for each animal, if applicable. Cage side observations included, but were not limited to changes in the skin and hair-coat, eyes, mucous membranes, respiratory, circulatory, autonomic and central nervous system, and sensori- as weil as somatomotor activity and behavior pattern. Particular attention was directed to observation of tremors,
convulsions, salivation, diarrhea, lethargy, somnolence and prostration. The time of death was recorded as precisely as possible, if applicable. Since these signs can only be assessed adequately in their home cages, no specific assessment was performed during exposure while animals were restrained. During the course of study, additional clinical observations which took into account the pattern of examination consistent with a Functional Observational Battery (FOB). Measurements were made in 5 rats/sex/group. Each rat was first observed in its home cage and then individually examined. The following reflexes were tested: visual placing response and grip strength on wire mesh (wire-mesh grid-gripping resistance of the animal to pull), abdominal muscle tone, corneal and pupillary reflexes, pinnal reflex, righting reflex, tail-pinch response, startle reflex with respect to behavioral changes stimulated by sounds (e.g. finger snapping) and touch (back).

General clinical chemistry tests were performed at the end of the 3 months exposure period and at the end of the recovery period on all animals. The terminal blood samples were obtained by cardiac puncture of the deeply anesthetized, non-fasted rats (Narcoren®; intraperitoneal injection). The blood required for the glucose determination was obtained during the last exposure week (after urine collection) from the caudal vein of non-fasted rats. Anticoagulant-coated tubes were used except for blood collected for to examine hemostasis endpoints where sodium citrate was used as anticoagulant.
Hematology: Hematrocit, Hemoglobin, Leukocytes, Erythrocytes, Mean corpuscular volume, Mean corpuscular hemoglobin concentration, Mean corpuscular hemoglobin, Thrombocyte count, Reticulocytes, Heinz Bodies, Leukocyte differential count (Lymphocytes, Granulocytes, Segmented neutrophils, Eosinophilic neutrophils, Basophils, Monocytes, Plasma cells, miscellaneous abnormal cell types).
Clinical Pathology:
Aspartate aminotransferase, optimized (ASAT), Alanine aminotransferase, optimized (ALAT), Glutamate dehydrogenase (GLDH), y-Glutamylaminotransferase (y-GT), Lactate dehydrogenase (LDH), Alkaline phosphatase (APh), Albumin, Bilirubin (total), Blood glucose, Calcium, Chloride, Cholesterol, Creatinine, Magnesium, Phosphate, Potassium, Sodium, Total protein, Triglycerides, Urea,
Prothrombin time (PT, Ouick value, "Hepato Quick").

General urinalysis was performed towards the end of the 3 month study and 4 week post-exposure periods on 10 animals/sex/group. The rats' urine was individually collected overnight (approximately 3 7 a.m.) during the last study week from animals in metabolism cages (with watering bottles, pulverized Food ad libitum). The construction of the urine collection equipment did not allow contamination of urine with water or chow. The collection containers were cooled to nominal 12 °C in order to prevent/suppress uncontrolled growth of microorganisms overnight.
The following parameters were evaluated semiquantitatively (SQ) or quantitatively (Q) in the collected urine: .
Sediment composition (SQ), Urine osmolality (Q), pH (SQ), Volume (Q), Protein (SQ), Glucose (SQ), Blood (SQ), Bilirubin (SQ), Urobilinogen (SQ), Ketone bodies (SQ).

Ophthalmic examinations were conducted by a laboratory animal veterinarian or assistant trained in ophthalmoscopic examinations. Eye examinations were performed prior to the first exposure and towards the end of the exposure period. For examinations, an indirect ophthalmoscope (HEINE) was used. Five to ten minutes prior to the examination, the pupils were dilated with mydriatic (STULLN®). Routine screening examinations included an examination of the anterior segment of the eye, the posterior segment of the eye and adnexal structures. Structures examined in the anterior segment of the eye will typically include the cornea, sclera, iris, pupiI, lens, aqueous, and anterior chamber. Structures examined in the posterior segment of the eye will typically include the vitreous, retina and optic disc. Examination of adnexal structures will typically include conjunctiva, eyelids and eyelashes.
Sacrifice and pathology:
The following organs were weighted at necropsy after exsanguination: Adrenal glands, Brain, Heart, Kidneys, Liver, Lung (incl. trachea), Lymph nodes lung associated (LALN), Ovaries, Spleen, Testes, Thymus.
No organ weight data were collected from animals found dead. Paired organs were weighted together.

All surviving rats were sacrificed at the end of the respective observation period using sodium pentobarbital as anaesthetic and complete exsanguination by heart puncture (Narcoren®; at least 120 mg/kg body weight, intraperitoneal injection). All rats, irrespective of the day of death, were given a gross-pathological examination. Consideration was given to performing a gross necropsy on animals as indicated by the nature of toxic effects, with particular reference to changes related to the respiratory tract. All gross pathological changes were recorded and evaluated.
The following tissues were collected at necropsy:
Adrenals, aorta, bone and bone marrow section (sternum), brain (cerebrum, cerebellum, pons/medulla), epididymides, esophagus, eyes with optic nerve, eyelids, extraorbital lacrimal glands, femur with knee joint, Harderian glands, head with (focus) bulbus olfactorius and nasal cavity, heart, intestine (duodenum, jejunum, ileum, cecum, colon, rectum), kidneys including pelvis, larynx, liver, lungs plus trachea and main bronchi (all lobes), lymph nodes (lung associated, mandibular, mesenterics, popliteal, mediastinal), mammary gland, muscle, ovaries with oviducts, pancreas, pharynx, pituitary gland, prostate, salivary glands, sciatic nerve, seminal vesicles (incl. coagulation glands), skeletal muscle, skin (flank), spinal cord (cervical, thoracal, lumbar), spleen, sternum, stomach, testes, thymus, thyroid gland, tongue, trachea, ureters, urinary bladder, uterus with cervix, ureters, vagina and tissues with macroscopic findings.


Tissues were fixed in 10 % neutral-buffered formalin (NBF), except of eyes, testes and the left and two thirds of the right kidneys (fixed in Davidson's fixative then transferred to 10% NBF), and bone marrow processed for cytology which were fixed in methanol (evaluated only if warrented by the histopathology findings or clinical pathology). Examinations utilized lavaged (right lobes) lungs (all sacrifices, except week 13) and non..Javaged lungs (week 13 only). While the bronchus to the left lung lobe was tied up to allow lavage and instillation fixation of the right lung lobes, the left lobe was dissected and used for trecer determinations. The Jung or lung lobes were Inflated at a pressure of approxirnately 20 cm H20 with 10% neutral-buffered formalin, and then fixed similarly by Immersion. The femur, sternum, knee joint and nasal turbinates were fixed in 10% NBF, and then placed in a decalcification solution and will follow
routine processing in the histopathology laboratory.
In addition to the normal stain, lung sections were counterstained with an additional stain for collagen fibers, if indicated by the findings of this study. The seIection of the most appropriate stain is to the discretion of the pathologist. Histopathology was performed on all tissue specimens shown above in the control and high-level exposure groups. The tissues of the respiratory tract (Larynx, lymph nodes, lung, lung associated, pharynx and trachea), olfactory bulb (head) and brain were examined in all exposure groups. At terminal sacrifice, 10 rats/sex/group were examined while iterim (8 weeks exposure) or postexposure sacrifices (recovery groups) during weeks 17, 16, 39 the focus was on the lung (6 male rats/group/sacrifice) of . Other groups (and/or tlssues) were evaluated al the discretlon of the clinical pathologist only If warranted by specific changes. Of note is that Masson-Trichrome stained sildes of lungs were additionally examined.

Other examinations:
The rectal (colonic) temperatures were measured at several time points shortly after cessation of exposure (within 1/2 hour of cessation of exposure) using a digital rectal probe (H. Sachs, March, Germany). Five rats/group/sex were examined in exposure weeks 0 (day 0), 57, and 84.

Sampies of bronchoalveolar lavage fluid were collected from the lungs of rats (six male rats/group and time point) during week 8, 13, 17, 26, and 39. During the first two time points, the lavage was performed 1 day after exposure. Within the acellular supernatant of BAL-fluid (BALF), several indicators of pulmonary damage were assessed: Lactate dehydrogenase (LOH), Alkaline phosphatase, soluble collagen, total protein, ß-N-Acetyl-glucosamidase (ß-NAG), y-Glutamyltransferase (y-GT), total number of lavaged cells (including the volume and diameter ), cytodifferentiation

Biological specimens (Ieft lung lobe and lung-associated Iymph nodes (LAL)) were collected for the determination of cobalt (Co) as marker of exposure (the Co content in Baytubes is 0.53%). While the bronchus to the right lung lobes was tied-off to allow instillation fixation, the left lobe was dissected, weighed, and asservated. Co in organs was determined during weeks 8, 13, 17, 26, and 39.


The determination of Cobalt in lung tissue prompted an unscheduled post-study activity to better characterize the MWCNT structure retained in the lung uslng
digested BAl cells from cytospins by scanning electron microscopy (SEM). Thls analysis was qualitative In nature to reveal the key structures retalned In the pool of alveolar macrophages (spherical partlcles vs. fibrillar or tubular break-up products).

All variables that are not dichotomous are described by sex, dosage group and date using appropriate measures of central tendency (mean, median) and general variability (standard deviation, in most instances also minimum, maximum).

For the statistical evaluation of samples drawn from continuously distributed random variates three types of statistical tests are used, the choice of the test being a function of prior knowledge obtained in former studies. Provided that the variate in question can be considered approximately normally distributed with equal variances across treatments, the Dunnett test is used, if heteroscedasticity appeared to be more likely a p value adjusted Welch test is applied. If the evidence based on experience with historical data indicates that the assumptions for a parametric analysis of variance cannot be maintained, distribution-free tests in lieu of ANOVA are carried out, i.e. the Kruskal-Wallis test followed by adjusted MWW tests (U tests) where appropriate. Global tests including more than two groups are performed by sex and date, i.e. each sex x date level defines a family of tests in the context of multiple comparison procedures. Within such a family, the experiment-wise error is controlled. If not otherwise noted, all pair-wise tests are two-sided comparisons.

Results and discussion

Results of examinations

Details on results:
All exposures were tolerated without test substance-induced mortality. Two male rats succumbed during the exposure period, possibly to immobolization-related activities.

As can be noticed from Table 2 there were some incidental findings in individual rats from all groups. They are not considered to be causally related to the lest substance due to the absence of any time-dependent exacerbation or concentration-dependence. Thus, all rats tolerated the exposures without any last sUbstance-specific clinical signs.

The examination of reflexes did not reveal any differences between the groups.

In comparison to the concurrent air control group, there was no evidence of a conclusive, toxicologically significant effect on body (rectal) temperatures at any exposure concentration. Additionally, the temperature measurements made on control animals demonstrate clearly that the animal restrainer had no apparent effect on the body temperature (normal body temperature of the rats: 37.5°C - 38.5°C).

There was no toxicologically consistent, i.e., concentration- or time-dependent effect on body weights up to and including 6 mg/m³. Statistical significant changes appeared to be related to an increase rather than a decrease in body weights (vs. air control). Accordingly, as far as significant changes were observed they are considered to be of no pathodiagnostic or prognostic relevance. This interpretation is also supported by the analysis of incremental body weight changes (body weight gain).

There was no consistent evidence of effected food consumption across the exposure groups.

There was no consistently affected water consumption throughout the exposure period considered to be of toxicological significance.

There were no conclusive concentration-dependent changes between control and dose groups. Isolated statistical significances are considered to be of no pathodiagnostic significance.

There were no concentration-dependent changes in any group.Isolated statistical significances are considered to be of no pathodiagnostic or prognostic significance.

- Urinalysis:
There were no effects considered to be of pathodiagnostic relevance.

Ophthalmology performed prior to the start of the study and towards the end of the study did not reveal any conclusive evidence of test substance-induced changes in the dioptric media or in the fundus.

Collectively, with the exception of increased lung and LALN (lung associated lymph node) weights there ware no significant changes in organ welghts or the organ-to-body weight or organ-to-brain weight ratios. Due to the somewhat inconsistent changes in body weights, especially in male rats, the organ weight-to-brain weight ratio is considered to be of highest prognostic value. Based on this rationale, concentration-dependent increases in lung and LALN weights occurred at 0.4 mg/m³ and above. Borderline, although statistically significant increased LALN weights occurred in males at 0.1 mg/m³.

After 8 and 13 weeks of exposure, the lungs of rats exposed at 6.0 and 1.5 mg/m³ showed black to gray discolorations, respectively. Similarly, the LALNs of these groups showed also gray discolorations. After 13 weeks, a discoloration of LALNs was already evident at 0.4 mg/m³. During the course of the 6 months recovery period these gray to dark discolorations occurred at 0.4 m/m³ and above. LALNs appeared to be enlarged at 1.5 mg/m³ and above. Discoloration of lung associated Iymph nodes was black in the two upper concentrations and gray in the third concentration. Only one rat from the high concentration exhibited enlarged lung associated Iymph nodes.

Sex-related differences were not observed. Therefore the focus was always on males because this sex was used for time-course analyses during a postexposure period of 6 months. During the interim sacrifice (8 weeks) hypercellularity of the epithelium at the bronchiolo-alveolar junction occurred with incrsasing intensity at 0.4 mg/m³ andabove. This hypercellularity was characterized by an influx of inflammatory cells, mainly Iymphocytes, as well as thickened bronchiolar epithelium and alveolar septae. At the 13-week sacrifice, the principal exposure-relatad lesions in the upper respiratory tract were characterized by goblet cell hyper- and/or metaplasia, eosinephilic globules, and focal turbinate remodeling (thickening of turbinate bone with increased activity of osteoblasts) at 1.5 and 6 mg/m³. Minimal eoslnophilic globules or goblet cell hyper-/metaplasia were already apparent in some rats exposed at 0.4 mg/m³. There was no evidence of particle translocation to or reactive changes In the bulbus olfactorius. Lesions surrounding the bronchiolo-alveolar region were characterized by thickening of epithelial cell layers and influx of Inflammatory cells at 1.5 and especially 6 mg/m³ with gradually less prominent changes at 0.4 mg/m³. Slight to moderate inflammation, focally also with granulomatous appearance, was observed in rats exposed at 6 mg/m³. At this exposure concentratlon, a time-dependent increase of a multi-focal bronchiolo-alveolar hyperplasia was evident. Increased interstitial collagen staining (Sirius-red) was apparent at 1.5 and 6 mg/m³. Focal areas of increased collagen staining adjacent to sites of increased particle deposition and Inflammatory infiltrates occurred already at³. The borderline nature of increased collagen but in the absence of evidence of lung remodeling at 0.4 mg/m³, was substantiated further by Masson-Trichrome staining. Thickening of the visceral pleura, due to inflammation, increased collagen staining, and inereased particle retention, was apparent at 1.5 and 6 mg/rn³. Black pigmentation of LALNs, accompanied by increased paracortical cellularity, and increased bronchus-associated lymphoid tissue (BAL T) eontaining black particleladen macrophages occurred at aJl particle-exposure levels. These changes gained statistical significance at 0.4 mg/m³ and above. Collectively, pulmonary lesions were related to particle-related inflammatory responses at the site of initial deposition and retention of MWCNT structures at 1.6 and 6.0 mg/m³. Minimal ehanges related to the phagocytosis of free particles ware already apparent at 0.4 mg/m³. In summary, substance-induced findings were not detected in any extrapulmonary organs/tissues (apart from the lung-assoclated Iymph nodes and extrathoracic airways). The osteosarcoma seen in one rat at 0.4 mg/m³ in the nasal turbinates was considered to be spontaneous and unrelated to the inflammatory response. Apart trom few particle-Iaden alveolar macrophages, adverse response to particlerelated histopathological changes were not apparent at 0.1 mg/m³. Despite different grading scores and staining principles, the overall picture appears to converge. While a slightly increased collagen deposition was demonstrated al 1.5 and 6 mg/m³, a transition of increased interstitial collagen was apparent at 0.4 mg/m³ in the absence of structural changes suggestive of lung remodeling. Notwithstanding the fact that the increased septaI collagen staining observed 1.5 and 6 mg/m³ is conventionally taken as early evidence of interstitial fibrosis, the absence of higher scores of such changes renders the interpretation difficult due to the presence of the ongolng
chronic inflammation.

Due to the black color of Baytubes phagocytized by cells, the results from cytodifferentiation was predominated by 'non-classifiable' cells (NCs). Nonetheless, it appears to be scientifically justified to assume that most of the NCs are alveolar macrophages containing MWCNT with variable appearance due to necrotic/apoptotic transformation. With regard to bronchoalveolar lavage the concentration-dependent increase of neutrophilic granulocytes (PMNs) and BAL-collagen appear to be most robust and sensitive to probe for concentration-effects. For this lead endpoint a significantly
increased and sustained recruitment of infammatory cells occurred at 0.4 mg/m³ and above. The pattern of changes related to PMNs was not at variance at the acellular endpoints, especially LDH, protein, y-GT, and soluble collagen. All endpoints measured showed an increase between exposure weeks 8 and 13, suggesting that the changes in BAL reflect the increased cumulative lung particle burden.
The magnitude of changes In BAL (acellular and cellular) was somewhat less pronounced at the end of the 6-months postexposure period when compared to the end-of-exposure (week 13) time point. However. especially at exposure levels equal to and above 1.5 mg/m³, the findlngs are conslstent with a sustalned inflammatory response. Progressive changes did not occur in any group. Hence, in regard to BAL endpoints. 0.1 mg/m³ constitutes a no-observed adverse effect level with borderline changes of unclear toxicological significance at 0.4 mg/m³ (equivocal increase of extravasated soluble collagen, protein, y-GT, and equivocal influx of PMN). y-GT is a marker of increased activity of macrophages and Clara cells. Therefore, the
increases in this endpoint appear to be reflective of homeostatic changes (particle phagocytosis, increased syntheses of surfactant).

Kinetic analyses appear to suggest that Baytubes have a high biopersistence in lungs at overload conditions (retention half-time t1/2 approx. 350 days). Retention half-times at 0.1 mg/m³ were in the range of the limit of quantification of the tracer Co. While the increase in the cumulative lung dose was proportional to the exposure concentration, a duration-of-exposure-related increase in cumulative lung doses was not observed between exposure waeks 8 to 13 at all exposure levels. However, such increase occurred in LALNs of rats exposed at 1.6 and 6 mg/m³. Likewise, the toxicological responses increased with increasing exposure duration. Co translocation into LALNs occurred in a concentration-dependent as well as timedependent manner at 1.6 and 6 mg/rn³. In the lung, only minimal time-related changes in Co-tissue concentrations ware apparent during the 6-months exposure period. Placing these data into a 'poorIy soluble particle perspective', then the respective cumulative particulate matter dose per lung appears to have reached frank overloading conditions at 1.6 mg/m³ and above.

Post-study digestion of cytospins and analysis by scanning electron microscopy (SEM/EDX), did not reveal structures of disintegrated particulate material in the nano-size range. Similarly, isolated fibrillar structures ware not found. A time-dependent change in particle-size was not found as weil during the 6-months postexposure period. This finding provides experimental evidence that the assemblages of Baytubes are stable, i.e ., they do not break down in nano-particulates or produce fibrils.

Effect levels

Dose descriptor:
Effect level:
0.1 mg/m³ air
Based on:
test mat.
Basis for effect level:
other: nasal and pulmonary inflammatory responses

Target system / organ toxicity

Critical effects observed:
not specified

Any other information on results incl. tables

Table 1: Summary of subchronic inhalation toxicity (6h/day for 13 weeks)

Group/ Sex Target concentration (mg/m3) Toxicological result  Onset of mortality
 1/male 0 0 / 6/ 40  ---
2/male 0.1 1/ 7/ 40 77 d
 3/male 0.4 0 / 6/ 40 --- 
4/male  1.5 1/ 9/ 40  71 d
5/male   6.0 0/ 4/ 40 ---

Toxicological results:

number of dead animals / number of animals with signs after cessation of exposure / number of animals exposed

Table 1: Summary of subchronic inhalation toxicity - continuation

Group/ Sex Target concentration (mg/m3) Toxicological result  Onset of mortality
 1/female 0 0 / 2/ 10  ---
2/female 0.1 0/ 2/ 10 ---
 3/female 0.4 0 / 1/ 10 --- 
4/female  1.5 0/ 2/ 10  ---
5/female  6.0 0/ 0/ 10 ---

Applicant's summary and conclusion

Executive summary:

In a subchronic toxicity study (OECD TG 413) 10 male and 10 female Wistar rats per dose group were nose-only exposed for 13 weeks (6 hours /day, 5 days/week) to a respirable solid aerosol of MWCNT.

Micronization of MWCNT was applied to increase the dustiness of the test material without destroying the assemblage structure of MWCNT. This was verified by characterization the MWCNT before and after micronization. A confimatory analysis was also made from the aerosolized MWCNT which confirmed that the composition and morphology of the aerosolized MWCNT were essentially identical with the non-dispersed/non-micronized MWCNT.

The animals were exposed to 0, 0.1, 0.4, 1.5 and 6 mg/m³. Additional 30 male rats were exposed in the air control and all exposure groups followed by a maximum recovery period of 6 months. An interim sacrifice group of 6 animals per air control and dose group was necropsied after 8 weeks. Recovery groups wre sacrificed 4, 13 and 26 weeks after the end of the exposure period, respectively.

All exposures were toIerated without effect. This means, conclusive changes in body waights, food and water consumption, or specific clinical findings did not occur. Reflexes and body temperatures ware indistinguishable between the groups. A concentration-dependent and sustained increase of absolute end relative lung weights occurred at 0.4 mg/m³ and above. The increased weights of the lung associated-Iymph-nodes (LALN) mirrored the respective changes in lung weights which is consistent with a concentration-dependent increase of the particle clearance via Iymphatic pathways. Equivocal changes LALN weights were already noticed at 0.1 mg/m³. With regard to bronchoalveolar lavage (BAL )endpoints, 0.1m/m³ is considered to be a no-observed adverse effect level (NOAEL). There was only mild attenuation in the severity of the inflammatory response probed by BAL at 0.4 mg/m³. However, all end points suggestive of pulmonary inflammation were statistically indistinguishable from the control at the end of the 6-moths postexposure period. Histopathology demonstrated a complementary picture to BAL-analysis. After the 13 -week exposure period, goblet cell hyper- and/or metaplasia. eosinophilic globules and focal turbinate remodeling (thickening of turbinate bone with increased activity of osteoblasts) occurred in the nasal cavities at 1.6 mg/m³ and above. Minimal eosinophilic globules or minimal goblet cell hyper-/metaplasia ware apparent with low

incidence in some rats at 0.4 mg/m³. In thebulbus oIfactoriusneither evidence of particle translocation nor reactive findings existed. In the lungs. black particle-Iaden alveolar macrophages were observed Macrophages had an enlarged and/or foamy appearance at hlgher exposure levels. Black particles were observed both in the alveoli as well in the interstitium. Hypercellularily of airway epithelial cells at the bronchiolo-alveolar junction, including an increased influx of inflammatory cells and septal thickening, occurred in a concentration-dependent manner at 0.4 mg/m³ and above. Slight to moderate inflammation, focally with granulomatous appearance, occurred in addition at 6 mg/m³. Sirius-red stained and Masson-Trichrome lungs showed increased interstitial collagen at 1.5 and 6.0 mg/m³. Focally increased collagen was also detected adjacent to areas of increased substance deposition and inflammatory infiltrates at 0.4 mg/m³. In LALNs black pigmentation occurred at 0.4 mg/m³ and above. LALNs showed increased cellularity of the paracortex. Thickening of the visceraI pleura, due to inflammation, collagen deposition, and increased black particle load, was apparent in rats exposed at 1.5 mg/m³and above. Black macrophages in the BAL T and increased BALT were observed at all particle-exposure levels. During the course of the 6 months recovery period, black pigment in alveolar macrophages was detectable in all substance exposed rats. Likewise, black macrophages were present in the BALT. Inflammatory findings of the lungs did not decreased in severity to any significant extent with increasing postexposure duration. A reactive, multifocal bronchiolo-alveolar hyperplasia was slightly increasing during the recovery period at 6 mg/m³; however, without consistent time-related trend. Apart from LALNs at no exposure concentration extrapulmonary toxicity was observed.

Collectively, it is concluded that 0.1 mg/m³ constitutes the NO(A)EL in this study based on nasal and pulmonary inflammatory responses (this means at sltes of predominant particle deposition).