British Journal ofIndustrial Medicine 1985;42:36-42

Morphometric and elemental microanalytical studiesof human lung in health and diseaseK A SIEGESMUND,' A FUNAHASHI,2 AND D E YORDE3

From the Departments ofAnatomy, ' Medicine,2 and ofAnatomy and Pathology,3 Medical College ofWisconsin, Milwaukee, Wisconsin 53226, USA

ABSTRACT Current methods for determining the fibrogenicity of substances are based on rela-tively long term exposures of animals to the substance and the evaluation of morphologicalchanges occurring in the lung. The use of inhalation chambers, which produce a more physiologi-cal environment, suffer from the need for particularly long exposure times (1-3 years). Thepresent study describes a technique using scanning electron microscopy, energy dispersiveanalysis, and a digitiser pad with a computer to evaluate the fibrogenicity of silica in cases ofknown exposure. Scanning electron micrographs taken from silicotic lungs were evaluated for thedegree of thickening (fibrosis) and the same areas were analysed for silicon content. Correlationsbetween silicon content and septal thickening were shown to be significant (p < 0.0001). Thestudy also describes the concentrations of elements found in normal lungs. The technique forestablishing correlation curves between elemental concentrations and septal thickening could beof value in determining the fibrogenicity of pure substances after short exposures in an environ-mental chamber.

One of the primary morphological changes observedin fibrotic lung disease due to inhalation of inorganicdust, such as silica, is a pronounced thickening of theinteralveolar septa. The thickening is associatedwith the deposition of large quantities of collagenfibres among the cells of the septa. It is thought thatthe fibres are laid down by one of two processes. A"fibrogenic' agent may act directly on the fibroblaststimulating it to produce large quantities of collagenor the cell may also be stimulated by substancesreleased from dying macrophages that have ingestedexcessive amounts of non-digestible substances.'-6In either case the agent responsible directly, or indi-rectly, for fibroblast stimulation is considered to befibrogenic.

Several methods are available for assessing thefibrogenicity of a substance. The haemolysis of redblood cells is one method that has been used,79although a recent study has shown that this is notalways predictive of fibrosis.'01' Other methodsinvolve the determination of the cytotoxicity of thesuspected substance.'2'4 This procedure alsoappears to correlate poorly with fibrosis and is not

Received 14 November 1984Accepted 23 January 1984

recommended as the sole method for determiningfibrogenicity." '5

In vivo methods for determining fibrogenicity ofsubstances are based on the instillation of sub-stances, such as silica, into the trachea of experimen-tal animals.'6-'8 After a period of 6-12 months thelungs are examined for either biochemical or his-tological alterations. Inhalation studies for deter-mining fibrogenicity, although more physiological,need long exposures in chambers (1-3 years) beforecurrent methods can detect significant changes in thelung.'9 20

Biochemical methods involve the assay of lunghydroxyproline which is considered to be indicativeof collagen sythesis.2123 Histological methods areconsidered the most reliable means of determiningwhether or not a substance is fibrogenic."24 It isgenerally agreed that the presence of thickenedsepta together with interstitial exudate and epithel-ial metaplasia are features characteristic of pulmo-nary fibrosis. A substance can be evaluated for itsfibrogenicity by exposing animals to known concen-trations of the pure substance and observing for anyof these changes. Both instillation techniques (6months-i year) and inhalation techniques (1-3years) require long periods of exposure beforesignificant changes can be observed in the lung.

36

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://oem

.bmj.com

/B

r J Ind Med: first published as 10.1136/oem

.42.1.36 on 1 January 1985. Dow

nloaded from

Morphometric and elemental microanalytical studies ofhuman lung in health and disease

The examination of human lungs from workersexposed to various substances does not always per-mit the establishment of a "cause and effect" rela-tion between a particular substance and fibrosis. Thedetection of a foreign element in the fibrotic lungtissue does not necessarily establish the aetiologicalrole of that element. If, however, a positive correla-tion is observed between the amount of substance inthe tissue and the degree of fibrosis then a possibleaetiological role for that substance may be sug-gested.The present study describes a procedure using

scanning electron microscopy and energy dispersiveanalysis together with morphometry to quantitatesmall areas of fibrosis with known concentrations ofelements found within the fibrotic areas. A doseresponse relationship may then be implicated bet-ween the foreign substance and the septal thicken-ing. In addition, normal levels for various elementsfound in lungs of individuals without evidence ofpulmonary disease were determined.

Materials and methods

SPECIMEN PREPARATIONLung tissue from normal controls and establishedcases of silicosis with interstitial fibrosis wereexamined in the scanning electron microscope(SEM) and analysed by energy dispersive x rayanalysis. The individuals chosen as normal controlshad no known clinical or histological evidence oflung disease. Sections were cut from paraffin blocks,stained with haematoxylin and eosin, and examinedin a light microscope. Thirty micrometer sections oftissue showing areas of varying degrees of fibrosiswere cut from these blocks. The sections weretreated with xylene to remove the paraffin andhydrated by passage through decreasing concentra-tions of alcohol and into distilled water. The sectionswere floated from the distilled water on to the sur-face of spectroscopically pure carbon planchets. Thesamples were coated with a 100 A layer of carbon byvapour deposition.

SEM EXAMINATIONThe planchets containing the sections wereexamined in a JSMU 35 scanning electron micro-scope. Areas of thickened septa, selected to be freeof any obvious vessels, pleura, bronchi, or noduleswere photographed at 100 x (fig 1) and an energydispersive x ray analysis was carried out. The areasphotographed and analysed were about 1 mm insize. A total of five areas was photographed andanalysed for each patient or normal controlexamined. A nuclear semiconductor 152 eV resolu-tion Si(Li) detector and a Tracor Northern NS880

Fig 1 Scanning electron micrograph ofa 30 micrometerthick section ofhuman lung. (x 100.)

computer base data handling system were used forthe analyses. The specimen detector distance was2-85 cm at 0 degree tilt and a takeoff angle of 37°.All spectra were stored on magnetic tape for futurereference. Each area was analysed for 100 secondsat a 25 kv operating potential and a specimen cur-rent of 3 x 10-1 amps.

DATA REDUCTION AND USE OFELEMENT/SULPHUR RATIOSThe relative abundance of each element was calcu-lated from an analysis of the counts generated for agiven element by comparison with the referencespectra of the materials containing known concen-trations of that element. Concentration ratios werecalculated and corrected for absorption, fluores-cence, and atomic number effects, assuming carbon

%. X M r-!J--- '.a v -l

Fig 2 Scanning electron micrograph ofhuman lungsection with transparent grid overlay (black lines) used toselect random X, Y coordinates.

37

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://oem

.bmj.com

/B

r J Ind Med: first published as 10.1136/oem

.42.1.36 on 1 January 1985. Dow

nloaded from

38

0

4567891011121314151617

Siegesmund, Funahashi, and Yorde

+ + + + 14 + 4 4 + + 4 . . + + 4 . . . . . . .+ . . . . . . . + . . . . . . . . . . . . . .

+ + + + + + + + + + . + + . + + + + + + 17 + +O + . . 4 +. +. +. +. . . . . . + + .+ .* . . . . . 16 13 . . . + + + + + + + + + + . .

+ + + + + + + . + . + 11 + + + . 18 + 20 . . . .

+0 + + 7 . . . . .+ . . .+ . .+ .+ .++ . + . + + . . . . . 15 + . . . . 0 . . . . .

+ + + + . . + + 1 . . . . . . . . . . . 0 + ++ + + 0 + +. . . + 0 0 . . . . . . . . 12 + +9 + + 3 . . . . . . . . . . . . + +. . . . ++ . + . + + + . . + +. . . . . . + + + + . .

+ . . + +. . + + + +. . . . . . . . 5 +. ++ + + 0 +. . . +. . +. + + 0 + 4 * + + 2+ + + + + + + + + + + + + + + + + + + + + + ++ . 0. . . . . + . . . . . + + .+. . .L . . . . . 4 + + + . + +. . . . 8 + + +_+_+ +_+ +_+ + 0_10 _+_+_+_+_+_ 0 + + 19 6 +0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Picture No 8622experiment Example

49.13 44A58 44.66 103-41106.21 109-45 130-47 50-9019.53 106-48 70-75 121-5627.76 36.47 57.26 68.78

183.84 121-23 104.19 69.4Fig 3 Cathode ray tube display ofgrid system with 20 individual measurementsofsepta. Numbers on grid indicate location ofeach measurement.

as a background. Because of the difficulties inherentin determining the exact amount of materialexamined, the data were presented as relative ele-mental concentrations by weight.The use of Si/S ratios was necessary because of the

structural characteristics of lung tissue and because

ttickness

of slight differences in section thickness. A previousstudy using regression analyses on sections of vary-ing thickness has shown that sulphur may be used asa reliable indicator of tissue mass, and that a Si/Sratio gave a more valid indication of silicon concen-tration than silicon alone.2526 Based on Si/S ratios

Si to sulphur ratio

Si Y= 459-305X+ C8-7089 Coff of det= 0-496711Enter the symbol of the element for the comparison (CR-end)

Fig 4 Correlation curve for silicon versus septal thickness.

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://oem

.bmj.com

/B

r J Ind Med: first published as 10.1136/oem

.42.1.36 on 1 January 1985. Dow

nloaded from

Morphometric and elemental microanalytical studies ofhuman lung in health and disease

Septalthickness

I+ +l

- 5U§w M to sp ratioMg to sulphur ratio

Mg Y= 115.78X+ 195-671 Coff of det=6.64882E-O0Enter the syrbol of the element for the opcriso (CR-end)

Fig 5 Correlaton curve for magnesium versus septal thickness.

obtained from energy dispersive x ray analysis ofinteralveolar septa this earlier study established aratio of below 0-2 as a normal level, 0O2 to 0 3 as araised level, and above 0-3 as a level of silicon that isconsistent with silicosis.

MORPHOMETRIC PROCEDUREA total of 13 cases of silicosis and five normal con-

Septal

trols was evaluated. All of the control lungsexamined were taken from necropsy specimens ofindividuals who had died of non-pulmonary causes.The silicotic lungs examined were taken from eightnecropsy and five biopsy specimens. The site fromwhich the necropsy tissue was taken was not known.All biopsy tissue was taken from the right middlelobe.

Al to suphur ratioAl Y= 1186.19K083.7133 Coff of det=0.531014Enter the symbol of the element for the comparison (CR-end)

Fig 6 Correlaton curve for aluminum versus septal thickness.

39

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://oem

.bmj.com

/B

r J Ind Med: first published as 10.1136/oem

.42.1.36 on 1 January 1985. Dow

nloaded from

40

Table 1 Correlation between elemental content and septalthickness

Element Correlation coefficient p Value ('J

Na 0-09 0-1Mg 0-03 0-1Al 0-73 0-0001Si 0-70 0-0001P 0-63 0-001Cl 0-05 0-1K 0-35 0-05Ca 0.15 0-1Mn 0-29 0 05Fe 0-31 0 05Cu 0-06 0 1

In order to determine septal thickness from thescanning electron micrographs a digitiser pad and aTalos smart box interfaced with a Cromemco Sys-tem III, 64k computer were used. Five scanningelectron micrographs were taken from each caseexamined. The average septal thickness in the areasanalysed was calculated by averaging the width(linear) of 20 septal measurements chosen at ran-dom by the computer using a transparent grid over-lay placed on the electron micrograph (fig 2). Acomputer program then randomly selected one ofthe X and one of the Y coordinates on the overlay. Ifthe coordinate intercepted a septum the width ofthat septum was measured at that point using thedigitising pad and the data were automaticallystored in the computer. If the coordinate interceptedan alveolar space then another coordinate wasselected until 20 measurements were attained. Fig-ure 3 shows the cathode ray tube display of the gridsystem with the 20 individual septal measurements.The average septal thickness of these 20 measure-ments was then calculated and compared directlywith the silicon concentration of the septa in thisparticular field. The computer program was used todetermine if a correlation existed between the ele-

Table 2 Concentration ofelements in normal lung at 100x magnification (element/S ratio x 10-3)

EIS units

Element Average + 3 SD

Mg 5 13Al 8 45Si 64 148P 672 1056S _ _Cl 34 91K 28 82Ca 362 680Mn 19 49Fe 192 630Cu 58 211Zn 48 153

Siegesmund, Funahashi, and Yorde

mental content of the septa and the septal thicken-ing in the same field. The coefficients of correlationswere calculated for each element found in the lung.Data were stored on two dual 8" floppy discs underthe appropriate label and viewed on a Tektronix4025 terminal with graphics capability. A Tektronix4652 hard copier and Diablo 1640 printer were usedfor hard copy.

Results

MORPHOMETRY MEASUREMENTSSeptal measurements were made using the talosdigitiser pad as described and the data stored in thecomputer. Eighteen cases (5 controls and 13 silico-tics) were examined. At least one area from eachcontrol or silicotic case was examined. In four casesof silicosis the extent of fibrosis was so severe that nouseful data could be obtained. An attempt was madeto photograph areas varying in degree from moder-ate to high regions of fibrosis. It is recognised thatthe absolute septal thickness of the septum cannotbe measured in 30 micrometer thick sections. Theaverage relative septal thickness of control lungs,however, using a grid system, was 16 ,um.The coefficients of correlation were calculated for

each element found in the silicotic lungs. A correla-tion of 0-71 (p < 0.0001) was found between Si andseptal thickening (fig 4). This correlation is based onmeasurements from 22 different areas of varyingdegrees of fibrosis. The positive slope of the curvesuggests that a good correlation exists between thesilicon content and the thickness of the septa.By comparison, the coefficient of correlation for

magnesium was 0-02. The graph of this correlation(fig 5) was extremely flat by comparison with silicon.The only other element with a high correlation wasaluminium (fig 6) where a coefficient of correlationof 0-73 was determined. Table 1 shows the correla-tions of 11 different elements with septal thickness.The p values for each element indicate thesignificance of the relationship of each element toseptal thickness.

Since both Al and Si showed a high correlationwith septal thickness, a correlation was determinedfor silicon after the removal of all the highaluminium cases. Only those cases with less thanthree standard deviations above the mean ofaluminium were included in the new calculations.After removal of the six high Al cases the correla-tion coefficient for Si and septal thickness remainedsignificant at 0 58 (p < 0.05). The correlationcoefficient of Al to septal thickness, however, wasonly 0-16 (not significant). This suggests that highaluminium levels are not necessary for the correla-tion between silicon and septal thickening.

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://oem

.bmj.com

/B

r J Ind Med: first published as 10.1136/oem

.42.1.36 on 1 January 1985. Dow

nloaded from

Morphometric and elemental microanalytical studies ofhuman lung in health and disease

CALCULATION OF NORMAL VALUES OFELEMENTS IN LUNGA selection was made of five random areas of lungfrom 20 normal individuals without evidence ofpulmonary disease. These areas were analysed andthe average values for all measurable elements cal-culated.The concentration of the elements were expressed

as ratios of the number of x ray counts of elementdivided by the x ray counts for sulphur. This com-

pensates for both section thickness and countingtime. The derived ratio was then multiplied by 10-3and expressed as EBS units (element/sulphur). Usingthis technique, for example, the average silicon con-

centration in lung would be a silicon/sulphur ratio of0-064 or 64 E/S units.

Table 2 shows the normal values for each elementdetected in the lung sections. The standard deviationof each element was calculated and a value of +3standard deviations above the norm was arbitrarilyselected as being raised for any particular element.As already stated, a Si value of 300 E/S units was

previously determined to be diagnostic for silicosis.Elements not listed in the table cannot be detectedby energy dispersive x ray analysis in lungs ata 10Ox magnification. Although some elementsforeign to the lung may be detectable below 5E/S units, only elements 10 E/S units or above are

normally reported as raised.

Discussion

The use of element/sulphur ratios to express theamount of substances in the lung was particularlyimportant since low power analyses of the lung showconsiderable variation in tissue mass for each area

analysed. The use of sulphur as an internal standardto account for tissue mass provides a more reliableindication of the actual elements present in a unitmass of septa.2526

It is recognised that the normal levels of elementsreported in the lung sections represent only- thoseelements which are not extracted through the prep-

aration procedures. It is also possible that organicsubstances, not detectable by this method, could bereported in the lung sections represent only thosethe method is the inability to identify the form of theelement in the human lung sections. The implicationthat a substance is producing fibrosis must, there-fore, be made with caution when using human mat-erial where there is no history of exposure to a par-

ticular substance.Although the cases studied here are well

documented cases of exposure to silica, it is usuallynot possible, without using individual particle

analysis, to distinguish between silicate and silica.The high (p < 0.001) correlation of phosphorus withseptal thickening was not unexpected since fibroticareas contain high concentrations of macrophagecells. Nuclei from dead macrophage cells probablycontribute phosphorus to the high readings. This hasbeen seen by us in the analysis of areas of high cellu-larity such as tumours that also show high phos-phorus content. Since phosphorus is not a foreignelement in tissue it is most likely not a cause offibrosis.The high correlation of silicon to septal thickening

or fibrosis was not surprising since many studieshave clearly shown that many forms of crystallinesilica can produce fibrosis.Many of the silicotics included in this study were

foundry workers who were also exposed to highlevels of iron. In view of this it is interesting to notethat although iron levels were high in some of thelungs, iron showed a poor correlation to the thick-ened septal areas. This is consistent with earlierstudies which showed iron and silicon deposited indifferent locations in lung septa.25The high correlation of aluminium to septal thick-

ening was interesting in view of several reports ofaluminium pneumoconiosis.2730 It should be noted,however, that all the high aluminium areas also con-tained moderate to high levels of silicon. Removal ofthe high aluminium areas still provided a good corre-lation of silicon to septal thickness but aluminium nolonger correlated to septal thickening (0-16). Thissuggests that low levels, at least of aluminium, donot produce significant septal thickening.Goralewski' s view that high levels of aluminium maycause fibrosis of the lung3' could not be tested sincenone of our high Al cases was without high levels ofsilicon.The technique described in this paper could be of

value in evaluating the fibrogenicity of pure sub-stances on animal lungs. The determination ofelement/fibrosis correlations provides a relativelysimple and rapid means of evaluating the fibrogenicpotential of substances. A correlative study may bepossible with short exposure times and with only amoderate thickening of the septa. This would be ofparticular value when working with animal modelswhere septa are usually thin. The early detection ofseptal thickening by this method would" make theuse of inhalation chambers more practical andwould provide for a more physiological environmentthan intratracheal injection. Since small increases inseptal thickness can be determined easily, therequirement for prolonged inhalation studies couldbe avoided. The use of inhalation chambers wouldalso prevent the uneven distribution of substancesamong the various lobes.

41

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://oem

.bmj.com

/B

r J Ind Med: first published as 10.1136/oem

.42.1.36 on 1 January 1985. Dow

nloaded from

42

References

Heppleston AG, Styles JA. Activity of a macrophage factor incollagen formation by silica. Nature 1967;214:521-2.

2 Heppleston AG. The fibrogenic action of silica. Br Med Bull1969; 25: 282-7.

3Kilroe-Smith TA, Webster I, Van Primmelen M, Marasas L. Aninsoluble fibrogenic factor in macrophages from guinea pigsexposed to silica. Environ Res 1973;6:298-305.

Aalto M, Potila M, Kulonen E. The effects of silica-treatedmacrophages on the synthesis of collagen and other proteins invitro. Exp Cell Res 1976;97: 193-202.

Aalto M, Kulonen E. Fractionation of connective-tissue-activating factors from the culture medium of silica-treatedmacrophages. Acta Pathol Microbiol Immunol Scand (C)1979;87:241-50.

6Burrell R, Anderson M. The induction of fibrogenesis by silica-treated alveolar macrophages. Environ Res 1973;6:389-94.

Hefner RE Jr, Gehring PJ. A comparison of the relative states ofhemolysis induced by various fibrogenic and nonfibrogenicparticles with washed rat erythrocytes in vitro. Am Ind HygAssoc J 1975;36:734-40.

8Ligacz J, Paradowski Z. Hemolytic activity of aluminum and itscompounds. Med Pr 1973;23:41-5.

Morgan A, Holmes A, Talbot RJ. The hemolytic activity of somefibrous amphiboles and its relation to their specific surfaceareas. Ann Occup Hyg 1977;20:39-48.

'° David A. Hemolysis in vitro in the study of fibrogenity of indus-trial dusts. Int Arch Occup Environ Health 1976;37:289-300.

David A, Hurych J, Effenbergerova E, Holusa R, Simecek J.Laboratory testing of biological activity of ore mine dust:fibrogenicity, cytotoxicity and hemolytic activity. Environ Res1982;24: 140-51.

12 Adamis Z, Timar M. Studies on the effect of quartz, bentoniteand coal dust mixtures on macrophages in vitro. Br J ExpPathol 1978;59:411-5.

'3 Harington JS, Allison AC, Badami DV. Mineral fibers: chemical,physiochemical and biological properties. Adv PharmacolChemother 1975; 12:291-402.

4 Schlipkotu HW, Beck EG. Observations on the relationshipbetween quartz cytotoxicity and fibrogenicity while testing thebiological activity of synthetic polymers. Med Law 1965;56:485-93.

'5 Kysela B, Robock K, Weiss Z, Skoda V. Relations of some phys-ical characteristics of quartz to fibrogenicity and cytotoxicity.In: 3rd symposium on experimental silicosis 1974. Ostrava:

Siegesmund, Funahashi, and YordeDum Techniky CVTS, 1974:5-17. (In Czechoslovakian.)

16 Dauber JH, Bossman MD, Pietra GG, Jimenez SA, Daniele RP.Morphologic and biochemical abnormalities produced byintratracheal instillation of quartz into guinea pig lungs. Am JPathol 1980;101:595-607.

71 Davies JM. The fibrogenic effects of mineral dusts injected intothe pleural cavity of mice. BrJ Exp Pathol 1972;53: 190-201.

18 Gro P, Villiers AJ, de Treville TP. Experimental silicosis. ArchPathol 1967;84:87-94.

'9 Heppleston AG, Wright NA, Stewart JA. Experimental alveolarlipoproteinosis following the inhalation of silica. J Pathol1970; 101: 293-307.

20 Burns CA, Zarkower A, Ferguson FG. Murine immunologicaland histological changes in response to chronic silica exposure.Environ Res 1980;21:298-307.

21 Juva K, Prockop DJ. Modified procedure for the assay of H3- orC'4- labeled hydroxyproline. Anal Biochem 1966; 15:77.

22 Richards RJ, Tetley TD, Hunt J. The biological reactivity ofcalcium silicate composites: in vivo studies. Environ Res1981;26:243-57.

23 Lewis DM, Burrell R. Induction of fibrogenesis by lungantibody-treated macrophages. Br J Ind Med 1976;33:25-8.

24 Dauber JH, Rossman MD, Pieha GG, Jimenez SA, Daniele RP.Experimental silicosis, morphologic and biochemical abnor-malities produced by intratracheal instillation of quartz intoguinea pig lungs. Am J Pathol 1980; 101:595-612.

25 Funahashi A, Siegesmund KA, Dragen RF, Pintar K. Energydispersive x-ray analysis in the study of pneumoconiosis. Br JInd Med 1977;34:95-101.

26 Siegesmund KA. Elemental content in alveolar septa in variouspneumoconioses. (SEM/1980/II, SEM Inc.) Chicago; AMFO'Hare, 1980:485-91.

27 Jordan JW. Pulmonary fibrosis in a worker using an aluminumpowder. Br J Ind Med 1961;18:21-3.

26 Mitchell J, Manning GB, Molyneux M, Lane RE. Pulmonaryfibrosis in workers exposed to finely powdered aluminum. Br JInd Med 1961; 18:10-20.

29 Vallyathan V, Bergerson WN, Robichaux PA, Craighead JE.Pulmonary fibrosis in an aluminum arc welder. Chest1982;81:372-4.

30 McLaughlin AIG, Kazantzis G, King E, Leare D, Porter RJ,Owen R. Pulmonary fibrosis and encephalopathy associatedwith the inhalation of aluminum dust. Br J Ind Med1962; 19:253-63.

31 Goralewski G. The aluminum lung: a new industrial disease. ZGesamte Inn Med 1947;2:665-73.

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://oem

.bmj.com

/B

r J Ind Med: first published as 10.1136/oem

.42.1.36 on 1 January 1985. Dow

nloaded from