Long-Term Exposure to Urban Particulate Matter on the Ocular Surface and the Incidence of Deleterious Changes in the Cornea, Conjunctiva and Retina in Rats

We investigated the time-dependent deleterious ocular changes induced by urban particulate matter (UPM) in vitro and in vivo. UPM treatment decreased human corneal epithelial cell migration and survival. Fluorescein scores were consistently increased by UPM application for 16 weeks. One week of rest at 2 or 4 weeks led to a recovery trend, whereas two weeks of rest at 8 weeks induced no change. UPM treatment decreased the tear film break-up time at 2 weeks, which was thereafter maintained until 16 weeks. No changes were found after periods of rest. UPM-treated eyes exhibited greater corneal epithelium thickness than normal eyes at 2 weeks, which recovered to normal at 4 and 8 weeks and was significantly decreased at 16 weeks. Apoptotic cell number in the epithelium was increased at 2 weeks, which remained constant except at 8 weeks. IL-6 expression in the cornea of the right eye continually increased for 16 weeks, and significant recovery was only observed at 8 weeks after 2 weeks of rest. Ocular pressure was significantly increased in the right eye at 12 and 16 weeks. Topical UPM application to the eye induced deleterious changes to various closely related parts of the eye.


INORGANIC CONSTITUENTS
Analytical Approach: In the investigations at NIST, instrumental neutron activation analysis (INAA) and neutron capture prompt gamma activation analysis (PGAA) were used to directly determine mass fraction values for the elements, except for mercury. Wavelength-dispersive X-ray fluorescence spectrometry (WDXRF) was used to determine specific count rates for selected elements in SRM 1648a. Additional measurements with photon activation analysis (PAA), proton-induced X-ray emission spectrometry (PIXE), solid-sample graphite furnace atomic absorption spectrometry (SS-GFAAS) and WDXRF were provided by collaborating laboratories. All assays were designed to establish comparability of values between those conventionally certified in SRM 1648, representing the parent material, and the SRM 1648a measured in small sample sizes. These measurements confirmed that the composition of the material had not changed in storage and that the measured values in SRM 1648 can be utilized for confirmation of the value assignment of SRM 1648a.
Mercury measurements were made using cold-vapor mercury generation coupled with inductively coupled-mass spectrometry (ICP-MS) isotope ratio measurements. Single subsamples (160 mg to 230 mg) were taken from each of eight bottles and spiked with 201 Hg followed by microwave digestion. The value assignment of the mass fraction of mercury is based on cold-vapor (ICP-MS).
Homogeneity Assessment: The homogeneity of SRM 1648a was assessed by analyzing samples of approximately 1 mg from bottles selected by stratified random sampling (as described in "Homogeneity Assessment for Inorganic Constituents"). Kurfürst homogeneity factors derived from the results confirmed that a smaller than 1 % relative heterogeneity component of the uncertainty in the results is encountered for most elements by selecting sample sizes of 5 mg or larger.

GC/MS (II):
The GC/MS (II) analyses were conducted on test portions of 180 mg to 300 mg from each of three bottles extracted using PFE with dichloromethane at 100 °C and 13.8 MPa (2000 psi), designated as GC/MS (IIa) and test portions of 180 mg to 300 mg from three bottles extracted using PFE with dichloromethane at 200 °C and 13.8 MPa, designated as GC/MS (IIb). The concentrated extracts were passed through a silica SPE cartridge and eluted with10 % (volume fraction) dichloromethane in hexane. The PAH fraction was then analyzed by GC/MS using a 0.25 mm i.d. × 60 m fused silica capillary column with a proprietary relatively nonpolar phase (0.25 μm film thickness; DB-XLB).

GC/MS (III):
For GC/MS (III), test portions of approximately 500 mg from each of six bottles were extracted using PFE with dichloromethane at 100 °C and 13.8 MPa. The samples were cleaned-up using automated SPE with 1.8 g alumina columns and eluting with 9 mL of 35 % (volume fraction) dichloromethane in hexane. The samples were analyzed by GC/MS with a 0.25 mm i.d. × 60 m fused silica capillary column containing a 50 % (mole fraction) phenyl-substituted methylpolysiloxane phase (0.25 μm film thickness; DB-17, Agilent Technologies).
For all of the GC/MS measurements described above, selected perdeuterated PAHs were added to the particulate matter prior to solvent extraction for use as internal standards for quantification purposes.
Homogeneity Assessment for PAHs: The homogeneity of SRM 1648a was assessed by analyzing duplicate test portions of 150 mg to 250 mg from 10 randomly selected bottles. Samples were extracted, processed, and analyzed as described above for the GC/MS (I). No statistically significant differences between bottles were observed for the PAH mass fractions at a 150 mg to 250 mg sample size.
Value Assignment for PAHs: Certified mass fraction values were assigned for PAHs in SRM 1648a (Table 2) based on analyses using Soxhlet extraction and PFE at both 100 °C and 200 °C when the values did not change based on the extraction method [2]. For some PAHs, PFE at higher temperatures resulted in higher levels of extraction for air particulate SRMs [3,4]. For these PAHs, the different mass fractions are reported as a reference values in Tables 5a and  5b and should be considered "method dependent".
Nitro-Substituted PAHs: SRM 1648a was analyzed at NIST for the determination of nitro-substituted PAHs. Duplicate test portions of 150 mg to 250 mg from 10 bottles were extracted using Soxhlet extraction for 20 h with dichloromethane. The concentrated extract was passed through an aminopropylsilane SPE cartridge and eluted with 20 % (volume fraction) dichloromethane in hexane. The concentrated eluant was then subjected to normal-phase liquid chromatography (LC) using a semi-preparative amino/cyano phase column with a mobile phase of 20 % (volume fraction) dichloromethane in hexane to isolate the nitro-PAH fraction. The nitro-PAH fraction was analyzed by GC with negative ion chemical ionization mass spectrometry (GC/NICI-MS) using a 0.25 mm i.d. × 60 m fused silica capillary column containing a 50 % phenyl-substituted methylpolysiloxane stationary phase (0.25 μm film thickness). Selected perdeuterated nitro-PAHs were added to the air particulate matter prior to extraction for use as internal standards for quantification purposes.
Value Assignment for Nitro PAHs: Nitro-PAHs were determined in SRM 1648a using only Soxhlet extraction, and as only a single extraction method was used, the values are provided as reference values ( Table 6). The reference values should also be considered as "method dependent" values because they are dependent on the extraction method and temperature. For the GC/MS and GC/NICI-MS analyses, two PCB congeners that are not significantly present in the air particulate extract (PCB 103 and PCB 198 [5,6]) and selected 13 C-labeled PCB congeners and pesticides were added to the air particulate material prior to extraction for use as internal standards for quantification purposes.

Certified Mass Fraction Values of Elements:
The certified mass fraction value for each element in Table 1, expressed on a dry-mass basis, is an equally weighted mean of the individual sets of results provided by the individual NIST methods and individual methods of the collaborating laboratories, where used. The NIST results and some of those provided by collaborating laboratories included estimates of all recognized sources of uncertainty. Some collaborating laboratory results were provided without complete uncertainty budgets. These uncertainties were augmented using an approach that accounts for the differences among the results obtained by different methods [7]. For values solely determined by NIST, the certified mass fraction are weighted means of the mass fractions from multiple analytical methods [8]. The uncertainty listed with each value is an expanded uncertainty about the mean [8,9], with coverage factor, k = 2, with an approximately 95 % confidence interval calculated according to the method described in the ISO/JCGM Guide [10,11]. The measurands are the total mass fraction for each element in urban particulate matter as listed in Table 1. Metrological traceability is to the SI derived units of mass fraction (expressed as mg/kg or percent on a dry-mass basis). 51.0

Certified Mass Fraction Values for PAHs:
The certified mass fraction values in Table 2 for the PAHs are weighted means of the mass fractions from multiple analytical methods [8]. The uncertainty listed with each value is an expanded uncertainty, U, calculated as U = kuc, with coverage factor, k = 2. The expanded uncertainty about the mean is determined by combining within method variances with a between method variance [12] following the ISO/JCGM Guide [10,11]. The measurands are the total mass fraction of each PAHs in urban particulate matter as listed in Table 2. Metrological traceability is to the SI derived units of mass fraction (expressed as mg/kg on a dry-mass basis).

Table 2. Certified Mass Fraction Values (Dry-Mass Basis) for PAHs in SRM 1648a
Mass Fractions (mg/kg)

Certified Mass Fraction Values for PCB Congeners:
The certified mass fraction values for the PCB Congeners in Table 3 are weighted means of the mass fractions from multiple analytical methods [8]. The uncertainty listed with each value is an expanded uncertainty, U, calculated as U = kuc, with coverage factor, k = 2. The expanded uncertainty about the mean is determined by combining within method variances with a between method variance [12] following the ISO/JCGM Guide [10,11]. The measurands are the total mass fraction for each PCB Congeners in urban particulate matter as listed in Table 3. Metrological traceability is to the SI derived units of mass fraction (expressed as μg/kg on a dry-mass basis).

Reference Mass Fraction Values for Elements:
Reference values are based on NIST results from one method for each reported element in Table 4. The results are validated by the values previously reported for SRM 1648 [13] since INAA and XRF procedures established equivalence of SRM 1648a with SRM 1648 based on previously certified values in the latter. The uncertainties of the NIST results were augmented [7] on the basis of the previously reported differences among the results obtained by different methods in SRM 1648 [13]. These results do not fulfill the criteria for certification since a full estimate of method bias for the determinations in SRM 1648a is not available. The reporting follows the ISO/JCGM Guide [10,11]. The measurands are the total mass fraction of elements in urban particulate matter as determined by the method listed in Table 4. Metrological traceability is to the SI derived units of mass fraction (expressed as mg/kg or percent on a dry-mass basis).   Table 5a and 5b based on the extraction method and conditions. Metrological traceability is to the SI derived units of mass fraction (expressed as mg/kg on a dry-mass basis). 0.186 ± 0.031 Acenaphthene (c,d) 0.250 ± 0.083 Acenaphthylene (c,d) 0.173 ± 0.012 Fluorene (b,c,d,g) 0.251 ± 0.035 Dibenzothiophene (b,c,d,g) 0.262 ± 0.015  (b) 0.098 ± 0.004 (i) Naphtho[1,2-b]fluoranthene (b) 0.635 ± 0.020 (i) Naphtho[2,3-b]fluoranthene (b) 0.186 ± 0.007 (i) Dibenzo[a,k]fluoranthene (b) 0.057 ± 0.007 (i) Dibenzo[j,l]fluoranthene (b) 0.518 ± 0.024 (i) Dibenzo[a,l]pyrene (b) 0.076 ± 0.007 (i) Naphtho[2,3-e]pyrene (b) 0.227 ± 0.004 (i) Naphtho[2,1-a]pyrene (b) 0.438 ± 0.017 (i) Dibenzo[e,l]pyrene (b) 0.355 ± 0.019 (i) (a) The reference mass fraction values, except where otherwise footnoted, are weighted means of the mass fractions from multiple analytical methods [8]. The uncertainty listed with each value is an expanded uncertainty, U, calculated as U = kuc, with coverage factor, k = 2. The expanded uncertainty about the mean is determined by combining within method variances with a between method variance [12] following the ISO/JCGM Guide [10,11]. (e) The reference value is a weighted mean of average mass fractions, with one average from each of two or more analytical methods [8,9]. The uncertainty listed with each value is an expanded uncertainty, U, calculated as U = kuc, with coverage factor, k = 2. The expanded uncertainty is the half width of a symmetric 95 % parametric bootstrap confidence interval [14], which is consistent with the ISO/JCGM Guide [10,11].  1.56 ± 0.07 Acenaphthene (b) 0.347 ± 0.022 Acenaphthylene (b) 0.722 ± 0.067 Fluorene (b) 0.406 ± 0.055 Dibenzothiophene (b) 0.606 ± 0.025 0.362 ± 0.055 4-Methylpyrene (b) 0.454 ± 0.075 Retene (b) 0.923 ± 0.032 Benzo[c]phenanthrene (b) 0.584 ± 0.067 Benzo  [8]. The uncertainty listed with each value is an expanded uncertainty, U, calculated as U = kuc, with coverage factor, k = 2. The expanded uncertainty about the mean is determined by combining within method variances with a between method variance [12] following the ISO/JCGM Guide [10,11].
The uncertainty listed with each value is an expanded uncertainty, U, calculated as U = kuc, with coverage factor, k = 2. The expanded uncertainty about the mean is calculated by combining within method variances with a between method variance [12] following the ISO/JCGM Guide [10,11]. The measurand in each case is the mass fraction for each analyte listed based on the method indicated. Metrological traceability is to the unit milligram analyte per kilogram sample on a dry-mass basis. (e) GC/MS (Vb) on a relatively non-polar phase after Soxhlet extraction with dichloromethane. (f) GC/MS (Va) on a 50 % phenyl-substituted methylpolysiloxane phase using the same extracts as GC/MS (Vb).

Information Mass Fraction Values for the Content of Selected Elements:
Information values that may be of interest and use to the SRM user are given in Table 9. Information values are based on results that did not allow complete assessment of all sources of uncertainty; hence, only estimated means without uncertainties are given. These element values deviate from values in the previous SRM 1648. Scandium (Sc) and thorium (Th) were found to be inhomogeneous at the 5 mg sample size.

SUPPLEMENTAL INFORMATION FOR SRM 1648a
Users may wish to refer to the compilation of literature data for the original SRM 1648 for further information on elements that may occur in the material and are not reported in this certificate [13].
Particle Size: Particle size distributions in SRM 1648a determined in aqueous suspension via laser light scattering instrumentation (Malvern Mastersizer 2000) and are shown in Figure 1. The suspensions were prepared by a 10 minute sonication in distilled water (20 mL with approximately 0.02 mg of particulate matter with a drop of 0.1 % solution [volume fraction] of Triton added). These suspensions were gradually introduced into the water-filled, small sample measurement cell until a 6.5 % obscuration of the laser beam was achieved. Each suspension was measured three times for 30 s with a 10 s pause between the passes. A refractive index of 1.52 and absorption index of 0.1 were selected for the measurements. Results were calculated using the General Purpose Model provided by Malvern. The results depicting a typical distribution for total suspended air particulate matter are shown in Figure 1. Uncertainties in these values are estimated at ± 10 % relative (2s). These values are provided for informational purposes only and are based on and instrument-specific measurement of SRM 1648a dispersed in water after 10 minute sonication in water. The values have not been confirmed by an independent analytical technique as required for certification. See Figure 1 for particle-size distribution. (b) d(0.5) is the particle-size distribution parameter indicating the particle size below which 50 % of the volume is present. (c) d(0.1) is the particle-size distribution parameter indicating the particle size below which 10 % of the volume is present. (d) d(0.9) is the particle-size distribution parameter indicating the particle size below which 90 % of the volume is present.  (Table 10)" for additional information.
Homogeneity Assessment for Inorganic Constituents: Three methods were used to investigate the homogeneity of SRM 1648a for the suggested sample size of several milligrams: small-sample INAA, SS-GFAAS, and WDXRF. Results as relative standard deviations for each of the certified elements are shown in Table 11. The contribution from heterogeneity uHET is derived from the measured total standard deviation uexp and its analytical contribution uAN according to the equation below (equation 1).
Based on the models linking sample mass (w) to the homogeneity of particulate materials [14], Kurfürst, Grobecker, and Stoeppler have proposed an elemental homogeneity factor He that gives the relative standard deviation in percent for the element of interest if 1 mg samples were repeatedly analyzed and no analytical uncertainty were to influence the result (equation 2) [15].
e HET H u w = (2) INAA Determinations: INAA has been shown to be applicable for the determination of heterogeneity in small samples because the small samples, which essentially form point sources, provide for great improvements in the assays [16]. In the case of this INAA procedure, the analytical variance is in many instances dominated by the uncertainty from counting statistics listed as "instrument uncertainty" in Table 11. Duplicate portions from 12 bottles and duplicate portions from one bottle containing samples from 6 randomly selected locations were analyzed by INAA, resulting in thirty-six test portions analyzed and the heterogeneity components were calculated by subtracting the analytical uncertainties from the observed experimental uncertainty.
WDXRF Determinations: WDXRF has been used routinely for homogeneity determinations because of the excellent instrument stability. Since the X-ray intensities are attenuated in the sample, the highest contribution comes from the surface sample layers. The analyzed sample mass varies for each element and was calculated from a sample thickness that contributes to 90 % of the X-ray yield. To obtain sample mass, the calculated thickness was multiplied with the sample area exposed to the excitation beam and multiplied with the sample density. The counting statistics are listed under "instrument uncertainty" in Table 11. WDXRF analyzed samples from 12 bottles in duplicate.

SS-GFAAS Determinations:
Solid sampling procedures were used in conjunction with GFAAS by directly weighing 20 µg test portions into the graphite furnace. The instrument uncertainty was determined as repeatability of the determination of 20 µg single-element solution standards. Twelve measurements were made for each test.