Immunologic and respiratory responses to airway challenges of dogs with toluene diisocyanate

MY PmU-um Chicago,

M.D., C. f3aymmd

20i#,

M.D., ud

KWWW

E. m,

B.S.,

III.

Dogs received diisocyanate

biweekly (TDI) that

or monthly aerosol immani:arion were cumulative doses anulogous

wirh single doses to hrrmun chronic

of tolrrene exposure

ut 0.02

ppm. The three dogs developed sysremic immune responses IO TDI-dog serum albumin (DSA). The immune responses included IgG. IgA . und IgU antibody responses and de,~elopmenr of’ lymphocyte reactiviry. IgA titers approached IgG N’ters and both were persistent, itshereas IgM titers were low and qf’ short duration. IgE antiho& titers demonstrated b! endpoint c’utuneorrs titration appeared but fluctuated und became negurive in spire of continued uerosol exposrrrc of TDI. Immediate-type airway responses occurred qfier TDl aerosolizution. Some limited responses that showed only ubnormulilies of selected pulmonuv firnclion puramrrers \t.crc cleurly not immunologically mediated becuuse they occltrred tc,irh initial e.cposure. Olhet immediute-type airwuy responses occurred that qualitatively simulated IgE-mediated. antigen-induced airway responses in dogs. ‘There was u s~aristicully signifcun, correlatron (p = 0.0268) between these larrer airway responses and immediate-?\pe skin reuc,ri\iry. This srud.v demonstrates that this dog model sima1utc.s rhe comple.xi!v ot’ aincqv responses to TDI in man btrr may proGde a luborurory system /or evaluation oJ’ immunologic, or other proposed

mechanisms of TDI asthma. (J ALLERGY

CUN

Inhalation of diisocyanates results in pulmonary responses in some workers exposed to these volatile agents. Although the pathogenesis of some of such responses may be toxic, other responses are immunologic in nature. A form of hypersensitivity pneumonitis and immunologic asthma due to MD1 occurs.’ We have recently reported that MD1 delivered to the bronchi of dogs results in multiple systemic immunologic responses, and a form of IgE-mediated asthma to MD1 can be induced.* The problem of human respiratory responses to TDI is complex. An excellent recent review of isocyanate respiratory disease places problems of TDI in perspective.3 When concentrations of TDI are

From

the Section

of Allergy-Immunology.

Department

of Medi-

cine, Northwestern University Medical School, Chicago, 111. Supported by U.S.P.H.S. Grant A! 11403 and Ernest S. Bazley Grant. Received for publication Oct. 5. 1982. Accepted for publication Jan. I I, 1983. Reprint requests: Roy Patterson, M.D.. Northwestern University Medical School, 303 E. Chicago Avenue, Chicago. IL 6061 I.

Vol.

71,

No.

6, pp.

604-611

IMUUNOL

71:6&t.

1983.

J

maintained below the levels that will produce tc#&& ity. asthmatic responses occur in some workers. S&n cases of TDI asthma have characteristics sugge&$ of an immunologic response to an inhaled the* acting as a hapten in a manner already d&Ineel$@ TMA asthma.4 The occurrence of asthmatic responses to TDI aposure at concentrations not considered to be tarnie led to the consideration that such reactions were timunologic in nature. Serologic studies led to repbb of IgE antibodies against TDI-protein coojuga@~L~ However, results of various laboratories have m vided data that vary in correlation of 1gE antiwith the presence of TDI asthma. The failure to dptplain TDI asthma in a totally satisfactory manner-by immunologic studies has led to alternative postof TDI action through direct receptor-stimula mechanisms or induction of bronchial byperreacti$& by unknown mechanisms.3 The problems of controlled studies of immune w sponses of workers exposed to immunizing chemirrJI resulted in animal studies of animal immune R sponses to TDI. These immunologic studies in s#+ mals have resulted in the demonstration of ant&qj@

VOLUME 71 NUMBER 6

Airway

Abbreviations used DSA: Dog serum albumin ELISA: Enzyme-linked.immunosorbent MDI: Diphenylmethane diisocyanate TDI: Toluene diisocyanate f: Breathing frequency PEFR: Peak expiratory flow rate V,: Tidal volume RL: Pulmonary resistance Cdyn: Dynamic compliance TMA: Trimellitic anhydride

challenges

assay

and lymphocyte responses against TDI-protein conjugates in various species.3 Clear correlations of IgE antibody with human TDI asthma have not been es-

tablished. Because of experience with animal immune responses to TMAj, 6 and MDI,2 using the immunizing chemical delivered to the airway, we decided to evaluate the responses of dogs to TDI, using meth-

odology that evaluates responses to TDI administered to the airway rather than immunization with TDIheterologous protein conjugates. By this system the chemical

reacts with airway

proteins

such

as DSA. Immune responses can be measured by various immunologic techniques using TDI-DSA as an antigen model of a TDI-haptenized airway protein. METHODS Animals Dogs were three healthy adult animals of mixed breeds, conditioned and observed to be free of acute or chronic respiratory tract disease and of parasites. Animals were anesthetized with pentobarbital prior to the various studies.

Reagents TDI was obtained from Aldrich Chemical Co., Inc., Milwaukee, Wise. DSA, chromatographically purified, was obtained from Cappel Laboratories, Cochranville, Pa. Rabbit anti-dog IgG, IgA, and IgM were purchased from Miles Laboratories, Inc., Elkhart, Ind. Alkaline phosphatase-conjugated goat anti-rabbit gamma globulin and pnitrophenyl phosphate substrate tablets (Sigma 104 phosphatase substrate) were obtained from Sigma Chemical Co., St. Louis, MO.

Preparation

toluene

diisocyanate

666

To document that TDI had conjugated with DSA, spectrographic analysis of the conjugate was undertaken and the molar ratio determined. The molecular ratio of TDI to DSA was determined as follows. Absorbance spectra were run. The differential between spectra was found to be maximal at 240 nm. Therefore the epitope density was calculated by using the difference in absorbance of TDI-DSA and DSA at equivalent DSA concentrations. The extinction coefficient of TDI was determined and was assumed to be the same in either the conjugated or unconjugated form. The molecular ratio of TDI to DSA was calculated and in the conjugates used ranged from 8 to 11.

Enzyme-linked

haptenizing

with

of TDI-DSA

The method of conjugation of TDI to DSA has been described previously .4 Briefly, 50 mg of TDI as a liquid was added dropwise to 10 ml of 9% NaHCO, containing 50 mg of DSA. After stirring for 30 min, the conjugate was dialyzed against four changes of 0. 1M NaHCO, and filtered through a 0.20 pm filter. As an additional control, DSA was taken through each of the above steps but without the addition of TDI. This was designated as sham-conjugated DSA.

immunosorbent

assay

The ELISA procedure was done according to methods previously described. ‘, * Briefly, Immulon MicroELISA plates of polystyrene (Dynatech Laboratories, Inc., Alexandria, Va.) were coated with antigen (0.175 mg/ml TDIDSA), 200 plliwell, in carbonate buffer (pH 9.6) and incubated overnight at 4” C. The plates were then washed three times with PBS-0.05% Tween (Sigma), 200 ~1 of dilutions of the dog sera was added to individual wells, and the plates were incubated for 2 hr at room temperature. Plates were washed three times with PBS-Tween, and 200 ~1 of a previously determined dilution of rabbit anti-dog IgG, IgA, or IgM was added to each well, and the plates were incubated at 4” C overnight. After washing again, 200 ~1 of alkaline phosphatase-conjugated goat anti-rabbit gamma globulin was added to each well in a previously determined optimal concentration and the plates were incubated 3 hr at room temperature. After a final washing to remove the unbound conjugate, the enzyme substrate p-nitrophenyl phosphate (Sigma 104 phosphatase substrate, 1 mg/ml) in 10% diethanolamine buffer, pH 9.8, was added and incubation was carried out for 30 min at room temperature. The enzyme reaction was stopped with 50 ~1 of 3M NaOH and the optical densities of each well were then read at 410 nm on a Dynatech MicroELISA Mini Reader MR 590. ELISA titer endpoints were read as the endpoint at which the test antiserum had an optical density reading at 410 nm that was 2 times that of the control serum.

Cutaneous

titrations

of TDI reactivity

Animals received 7.5 ml of 0.5% Evans’ blue dye intravenously. One-tenth milliliter of serial IO-fold dilutions of TDI-DSA (3.5 mg/ml) was injected intracutaneously. The strongest concentration had been shown not to be reactive in normal dogs. Positive reactions were deep blue areas at least 10 mm in diameter occurring within 30 min. Negative reactions showed no bluing. The titer was expressed as the lowest concentration of TDI-DSA giving a definite bluing reaction. Skin testing was also done with the shamconjugated DSA.

Immunization

procedures

Dogs were immunized with TDI every 2 wk for 4 mo and every 4 wk thereafter for 6 mo, with the following procedure. The animal was anesthetized and a cuffed endotra-

Dog Dog Dog

53?~ Sh-.a 662a---n

Weeks FIG. 1. IgG antibody aerosolized TDI. *Dog

cheal tube was inserted. aerosol, delivered with

response 549 died

Each dog received an in-line nebulizer

against after the

TDI-DSA demonstrated sixth challenge.

I mg/kg TDI by in a Bird Mark

VII (Bird Corp., Palm Springs, Calif.). It had been determined previously that in this system approximately 20% + 10% of the material aerosolized is retained in the airway.g For these challenges the number of breaths and concentration were varied to achieve this percent. The immunization also constituted an airway challenge in subsequent experiments. The amount of TDI to which a worker might be exposed at 0.02 ppm, the maximum recommended threshold concentrationrO over a 2 wk period, approximates 0.28 mg/kg for light work or 0.42 for heavy work. The aerosol dose for dogs for 2 wk is thus about two to three times that of a worker. After 4 mo of the biweekly exposure, two dogs received 1 mglkg every 4 wk. which approximates the 4 wk exposure of a human at heavy work. Pulmonary functions were monitored at this time. Skin titers and blood samples were obtained prior to each immunization.

These studies used previously described techniques developed to measure acute antigen-induced airway reactions in dogs or monkeys. “-‘.I Physiologic measurements determined f, PEFR, V,, RI,, and Cdyn: results were expressed as percent change from baseline established prior to the introduction of TDI into the airway.

LymphwyWtsadmmWn Lymphocytes were separated by a standard FicollHypaque method.r4 Reactivity to TDI-DSA was assayed using 3H-thymidine incorporation in a microculture technique.‘” Cultures were done in triplicate and stimulated with 50, 200, and 800 pglml TDI-DSA and 25 pg/ml phytohemagglutinin (as a positive stimulation control) and incubated for 4 and 7 days.

by

ELlSA

in dogs

receiving

RESULTS ~urnnrprrrw:b6,(eA.W~ After biweekly aerosol exposure to 1 mg/kg TDI, all three dogs developed systemic immune response. This was demonstrated by the ELBA using TDI-DSA as the antigen. IgG antibody response against TDIDSA is shown in Fig. 1. Under the conditions of aerosol exposure, a detectable antibody response occurred in 2 to 4 wk in all three dogs, and antibody levels persisted during the 4 mo biweekly TM aemsol exposure and during the subsequent TDI exposure given every 4 wk in the two dogs under continuous study

(Fig.

1). Of

note

is that

one

dog

witb

the

highest

early antibody response had a subsequent decline in IgG anti-TDI-DSA titers after monthly TDI exposure began whereas the dog with lower titers developed progressively higher titers after monthly exposure began (Fig. 1). One dog died after the sixth exposure to aerosolized TD1. IgA and IgM systemic antibody responses against TDI-DSA were also demonstrated (Fig. 2). These were persistent in relation to the TDI aerosol exposure. Although the IgA titers were less than the IgG antibody titers (Fig. I), in one dog, IgA anti-TDIDSA titer was only I dilution less than IgG titers. IgA antibody titers in dog 562 peaked before IgG antibody titers (Figs. 1 and 2).

IgE antibody response had to be assayed by immediate-type cutaneous reactivity to TDWSA because no reliable in vitro assay for camne IgE antibody against TDI-DSA is availseie. The concertbra-

Airway

VOLUME 71 NUMBER 6

1260

1

5

9

13

17

21

25

with

toluene

diisocyanate

607

l&l A, . . . ....fi......A ,:' f

A. '.., /A._........ \ ..!

1

challenges

29

33

37

o-o m--8

DOQ

537

Dog

549*

A-A

DOQ

562

41

Weeks FIG. 2. Systemic antibody response. B, IgM antibody

responses response.

against TDI-DSA *Dog 549 died

demonstrated after the sixth

c-4 I 0

by ELISA. A, IgA antibody aerosol TDI challenge.

A

o---o

Dog

537

A-.-A

Dog

562

Weeks FIG. 3. Cutaneous with 1 mglkg TDI.

reactivity

by endpoint

titration

tion of test antigen at 0.35 mg did not give a nonspecific irritant reaction in normal dog skin. Sham-conjugated DSA did not give a reaction in normal or sensitized dogs. Immediate-type skin reactivity appeared within 4 wk in all three dogs and was similar in titer. Results in two dogs over 41 weeks of study

of two

dogs

after

repeated

aerosol

challenge

are shown in Fig. 3. Both persistence and fluctuation of degree of cutaneous reactivity are shown. Immediate-type

airway

responses

For comparison, an Ascaris-induced, immediatetype airway response is shown in Fig. 4, A. This

J ALLERGV Ctltd. mn#tucw. J&E+33

n--d

R,

-"T

-

02

5

10

15

20

25

PEFR Cdyn

30

02

Time FIB. 4. Immediate-type mal pulmonary function parameters.

,o

5

15

20

25

30

five

abnorfunction

- Minutes

airway responses in dogs. A, Ascaris parameters. B, TDI response with

antigen response with five abnormal pulmonary

250

W-f 0-0 RL 04 PEFR o--o v, Cdyn

100 50 25 0 -25 -50

Time-Minutes FIG. 5. Airway responses occurring after initial TDI or subsequent TDI aerosol challenge. A, One abnormal pulmonary function parameter. 8, Two abnormal pulmonary function parameters. C, Three abnormal pulmonary function parameters.

model of asthma characteristically showed a response within 5 min, with an increase in f and RI, and a decrease in Cdvn, PEPR, and VT.‘* All pulmonary function parameters, particularly R,, may not be positive in each challenge, and previous extensive studies in dogs and monkeysa* I6 resulted in the selection of either four or five abnormal pulmonary function parameters and degree of abnormality of pulmo-

nary function parameters to classify a response as an immunologic, &E-mediated immediate-type airway response. In normal dogs, a similar aerosol challenge with I mg/kg TDI did not result in abnormal puimonary function parameters consistent with an immediatetype airway response by the above definition. In the dogs that had received aerosol immtmiz&on with

Airway

VOLUME 71 NUMBER 6

Pulmonary Function (PF): Frequency Abnormality: FIG. 6. Incidence

f

PEFR

V,

81%

55%

58%

challenges

with

toluene

diisocyanate

606

Cdyn

%

of

of abnormal

pulmonary

function

parameters

29%

13%

in 38 TDI

aerosol

challenges.

o--o

Dog

537

o--a

Dog

562

Weeks FIG. 7. Stimulation stimulation index

indices of lymphocytes in two of 2 or greater was considered

TDI, airway responses occurred which showed four to five abnormalities of pulmonary function parameters suggestive of an antigen-induced response. Fig. 4, B, shows a representative TDI airway response with five abnormal pulmonary function parameters. The pulmonary responses consistent with an antigen-induced response (four or five abnormal pulmonary function parameters) occurred in 10 of 38 TDI challenges, and in eight of 10 of these responses there was a positive cutaneous reaction to TDI-DSA (see Fig. 3). These airway responses occurred in all three dogs. Statistical examination of these results demonstrated that the positive respiratory responses correlated with positive cutaneous responses, yielding a chi-squared value of 4.9060, p = 0.0268. Analyses were done with a Hewlett-Packard HP85 computer. None of these responses occurred prior to the fourth TDI aerosol immunization. The TDI airway response shown in Fig. 4, B, with five abnormal pulmonary function parameters occurred three times in 38 TDI challenges. These were at peak cutaneous titers to TDI-DSA.

dogs receiving positive.

Other airway

aerosol

challenges

with

TDI.

A

responses

After some initial and subsequent TDI aerosol challenges we observed that abnormalities of selected pulmonary function parameters occurred which were not consistent with an antigen-induced airway response. These responses, which were apparently nonimmunologic, showed abnormal pulmonary function responses in one, two, or three parameters as shown in Fig. 5. The incidence of the different abnormal pulmonary function parameters and the degree of abnormality in 38 challenges is shown in Fig. 6. The effect on f, PEFR, and VT were most consistently abnormal. Lymphocyte

reactivity

Peripheral blood lymphocyte responses to TDIDSA occurred after initiation of TDI aerosol challenges. Results are shown in Fig. 7. A significant stimulation index in this laboratory is 2.0. Initial lymphocyte reactivity rose in the first few weeks after aerosol challenge with TDI, then subsided, but rose

again after monthly rather than biweekly TDl challenge began. After 16 TDI aerosol challenges, two dogs showed no clinical abnormalities other than the acute airway responses described above. One dog died alter the sixth TDI aerosol challenge and the type of airway response shown in Fig. 4, B. This dog expired within 24 hr and the lungs showed patchy hemorrhagic lesions. Histologic examinations* of the lung showed extreme capillary congestion, intra-alveolar capillary congestion, intra-alveolar hemorrhage, and bronchopneumonia. Many of the distal bronchi were filled with acute inflammatory cells, which extended into surrounding respiratory bronchioles and alveoli. No microorganisms were found in the lung sections.

Methodology used in these studies was derived from both human and animal studies evaluating the airway response to TMA. ‘* ‘* 6* *’ In the use of TMA in industry, four types of respiratory reactions, three immunologic and one irritant, occur.*8 The TMA reacts with airway proteins to produce TMA-human protein conjugates that immunize the host, and various immunologic techniques can be used to assess human and animal antibody and lymphocyte responses against TMA-self protein. These techniques proved applicable in the development of a dog model of MD1 immunologic lung disease2 and were then applied to determine whether information could be developed to provide a dog model of TDl on any aspect of TDI airway responses in man. Aerosol&d TDI was used rather than immunization with TDI-heterologous proteins to more closely approximate human exposure conditions, and the biweekly and then monthly exposures approximated concentrations inhaled by some humans at 0.02 ppm during the same period of time. An obvious difference is that the TDI aerosolized to dogs was in a single challenge rather than in a long-term workplace exposure. The dog model described might relate better to a high-dose short-term exposure such as might occur in a “spill” of TDI in the workplace. In a different species, TDI has been delivered by inhalation to guinea pigs, with an immune response demonstrated by precipitating antibodiis to TDI-guinea pig albumin or by passive cutaneous skin reactions.‘e The results demonstrate that under the conditions of

*Histologic examinations were done by Dr. David ment of PatMogy, Baylor University, Houston,

Yawn, Texas.

Depart-

these experiments an imtRtme ltz3ponse oc4xmd to TDI-DSA. We presume that this is the result of ‘I’D1 combining with DSA in the airway of the dogs aerosolized with TDI and immunization of the dogs with TDI-DSA. The immune response was demonstrated in four immunoglobulin classes, with the IgA titers approaching the IgG titers. The IgE antibody levels as demonstrated by immediate-type endpoint skin reactivity fluctuated more widely than in our past experience with dog IgE antibody against protein antigens resulting from aerosol exposure. This difference from the usual long duration of IgE antibodies in this species induced by airway exposure is unexplained and may suggest a difference in control mechanisms of IgE antibody synthesis with this antigen. Similarly, the lymphocyte reactivity was not persistently present against TDI-DSA. The airway responsiveness to TDI aerosol exposure in the concentration of 1 mg/kg demonstrated that this concentration of TDI resulted in immediate-type airway responses that were not immunologic in nature. Evidence for this is that two of three dogs responded at the time of the first aerosol immunizing exposure. These responses and various subsequent responses did not meet the criteria for a positive Q&mediated antigen-induced response, previously defined as four or five pulmonary function parameters demonstrating significant abnormalities. These limited responses may have an irritant or pharmacologic basis as has been suggested for some human TDI responses.” Other TDI-induced airway responses met the criteria for an immediate-type IgE-mediated airway response (Fig. 5). These responses occurred after prior aemsol exposure to TDI and usually in the presence of immediate-type skin reactivity. Two of 10 of these responses occurred in the absence of skin reactivity. Explanations for this may be that local pulmonary immune response may not be reflected in cutaneous reactivity at the time of these responses. In other studies of persisting systemic IgE-mediated mactivity of dogs to Ascan’s antigen, occasional negative ski tests occurred, apparently for technical reasons. Altematively, these (and perhaps other) antigen-type airway responses to TDI in dogs am not due to Igl$tntibody. This study shows that dogs exposed to- aerosolized TDI may be used to evaluate immune responses to TDI-dog airway proteins and that immediate-type airway responses occur or can be induced, which may be used to evaluate mechanisms of TDI airway response under controHed laboratory conditions.

I. Zeiss CR. Kanellakes TM, JJ, Pstternon R: hrunumgtih

BeBone JD, Levi@ D, RW&LY E-me&ted &ha and

hy-

VOLUME 71 NUMBER 6

2.

3. 4.

5.

6.

7.

8.

9. 10.

persensitivity pneumonitis with precipitating anti-hapten antibodies due to diphenylmethane diisocyante (MDI) exposure. J ALLERGY CLIN IMMUNOL6!5:346, 1980. Patterson R, Harris KE, Pruzansky JJ, Zeiss CR: An animal model of occupational immunologic asthma due to diphenylmethane diisocyanate with multiple systemic immunologic responses. J Lab Clin Med 99:615, 1982. Bernstein IL: Isocyanate-induced pulmonary diseases: a current perspective. J ALLERGY CLIN IMMUNOL 70~25, 1982. Zeiss CR, Patterson R, Pruzansky JJ, Miller MM, Rosenberg M, Levitz D: Trimellitic anhydride induced syndromes: clinical and immunological studies. J ALLERGY CLIN IMMUNOL 60:96, 1977. Patterson R, Roberts M, Harris KE, Levitz D, Zeiss CR: Pulmonary and systemic immune responses of rhesus monkeys to intrabronchial administration of trimellitic anhydride. Clin lmmunol Immunopathol 15:357, 1980. Sale SR, Patterson R, Zeiss CR, Fiore M, Harris KE, Yawn D: Immune response of dogs and rabbits to intrabronchial trimellitic anhydride. Int Arch Allergy Appl Immunol67:329, 1982. Voller A, Bidwell DE, Bartlett A: Enzyme immunoassays in diagnostic medicine: therapy and practice. Bull WHO 53:55, 1976. Sepulveda R, Longbottom JL, Pepys J: Enzyme linked immunosorbent assay (ELISA) for IgG and IgE antibodies to protein and polysaccharide antigens ofAspergillus,fumigurus. Clin Allergy 9:359, 1979. Patterson R, Harris KE: Aerosolized antigen dose-response studies in asthmatic monkeys. J Lab Clin Med 92:283, 1978. Verschueren K: Handbook of environmental data on organic chemicals. New York, 1977, Van Nostrand Reinhold Co., p. 596.

Airway

challenges

with

toluene

diisocyanate

611

11. Kelly JF, Cugell DW, Patterson R, Harris KE: Acute airway obstruction in rhesus monkeys induced by pharmacologic and immunologic stimuli. J Lab Clin Med S&738, 1974. 12. Patterson R, Mellies CJ, Kelly JF, Harris KE: Airway responses of dogs with ragweed and Ascaris hypersensitivity. Chest 65:488, 1974. 13. Patterson R, Harris KE, Clreenberger PA: Effect of prostaglandin D, and I, on the airways of rhesus monkeys. J ALLERGY CLIN IMMUNOL65:269, 1980. 14. Thorsby E, Brattie A: A rapid method for preparation of pure lymphocyte suspensions. In Terasaki PI, editor: Histocompatability testing. Copenhagen, 1970, Ejnar Munksgaard Forlag, p. 655. 15. Oppenheim JJ, Schecter B: Lymphocyte transformation. In Rose NR, Friedman H, editors: Manual of clinical immunology. Washington, D.C., 1976, American Society for Microbiology, p. 81. 16. Patterson R, Harris KE, Suszko IM, Roberts M: Reagin mediated asthma in rhesus monkeys and relation to bronchial cell histamine release and airway reactivity to carbocholine. J Clin Invest 57:586, 1976. 17. Patterson R, Zeiss CR, Roberts M, Pmzansky JJ, Wolkonsky P, Chacon R: Human antihapten antibodies in trimellitic anhydride inhalation reactions. J Clin Invest 6297 1, 1978. 18. Zeiss CR, Wolkonsky P, Ruzansky JJ, Patterson R: Clinical and immunologic evaluation of trimellitic anhydride workers in multiple industrial settings. J ALLERGY CLIN IMMUNOL 70: 15, 1982. 19. Karol MH, Dixon C, Brady M, Alarie Y: Immunologic sensitization and pulmonary hypersensitivity by repeated inhalation of aromatic isocyanates. Toxic01 Appl Pharmacol53:260, 1980.