Effects of Pesticides on the Immune Response Walter B. Dandliker", Arthur N. Hicks, Stuart A. Levison, Kris Stewart, and R. James Brawn Department of Biochemistry, Scripps Clinic and Research Foundation, La Jolla, Calif. 92037
T h e influence of various pesticides on the humoral and cellular immune response to fluorescein-labeled ovalbumin has been analyzed. Pesticides (Aroclor 1260, Dinoseb, Parathion, pentachloronitrobenzene, piperonyl butoxide, mixed pyrethrins, and Resmethrin) were administered intragastrically in corn oil in one dose (one-half of LIlio) before primary immunization. Control groups included those treated with corn oil alone or immunosuppressed with Methotrexate. Booster immunizations and test bleedings were scheduled a t weekly intervals thereafter. The cellular immune response was quantified by redness and swelling, histological examination, and by differential temperature measurements of the foot pads after antigen challenge. T h e concentration, binding affinity, and heterogeneity of the serum antibody were determined by fluorescence polarization measurements. Dinoseb and Parathion depress both the humoral and cellular response. Methotrexate and pentachloronitrobenzene give a late stimulation, while Resmethrin gives an early, sometimes very marked stimulation of the cellular immune response. Other pesticides showed little or no effect under the conditions tested. Effects on the humoral response were limited to changes in antibody concentration, the binding affinity being nearly constant in all instances. T h e world-wide use of pesticides makes it urgent to know as much as possible about the effects of pesticides and their degradation products on humans and animals. Immediate toxic effects are relatively readily assessed, but slow or delayed effects are more difficult to detect and yet may be the more important-possibly leading to altered susceptibility to disease, damage in utero, accelerated aging, etc. T h e immune response offers one parameter that is a sensitive indicator of a variety of physiological functions and which is readily quantified in a number of ways. The early work in this area was reviewed by Ercegovich ( I ) . Since that time there has been a growing interest in evaluating pesticide effects on a number of different aspects of the immune response. These include effects on antibody production, dermal reactions to specific immunogens, immunoglobulin levels, resistance to infection, complement levels, lymphoid cell counts, and effects on lymphoid organs detected histologically. T h e parameters most easily quantified are those connected with the humoral response, since serum antibody can be easily obtained and can be characterized in a number of ways. T h e production of anti-human serum albumin was found to be inhibited in rats injected with Lindane, 60 or 120 mg/kg ( 2 ) , while the titer of anti-bovine serum albumin was not consistently altered in chicks fed mash containing u p to 625 ppm of D D T ( 3 ) . Hemagglutinin levels of rats immunized against sheep red blood cells were suppressed by an oral dose of Methylnitrophos or Chlorophos, 5 or 7 mg kg-' day-', especially in rats fed a protein-deficient diet ( 4 ) .A second class of specific immunogen used to monitor the humoral immune response in pesticide-treated animals includes bacteria and viruses. Significantly lowered titers of anti-Salmonella t i p h i were found in rabbits given drinking water containing 300 ppm of D D T ( 5 - 7 ) . Also, the expected increase in the yglobulin 7s fraction in response to Salmonella inoculation was inhibited by Dieldrin and benzene hexachloride (8).On the other hand, no consistent effect was found in the anti-Salmonella pullorum titer of DDT-fed chicks ( 3 ) .Similarly, no 204
Environmental Science & Technology
differences in bacterial agglutination, indirect hemagglutination, indirect hemolysis, or precipitation were noted between Warfarin-treated and untreated rabbits immunized with purified Salmonella typhi endotoxin (9). Lower titers of tetanus antitoxin were consistently found when animals immunized with tetanus toxoid were given dietary Aroclor 1260 or Clophen A60 (50 ppm) ( I O ) or 0.1-0.2 LDso of Carbaryl orally ( II ) . Effects on antibody-mediated immunity have been investigated for Anthio and Milbex in goats (22, 13) and for Minex and D D T in chickens ( 1 4 ) .Examination of lymphoid organs proved to be a sensitive indicator of immunosuppression in a study of the effects of DDT, Aroclor 1254, Carbaryl, Carbofuran, and Methyl Parathion in rabbits (15). T h e present study was designed to simultaneously analyze the influence of various pesticides on the humoral and cellular immune responses to a well-defined immunogen (fluorescein-labeled ovalbumin). T h e pesticides were administered in one oral dose preceding primary immunization. Booster immunizations were then given periodically after sampling the serum and performing in vivo tests of cellular immunity. These included visual evaluation of redness and swelling of a challenged footpad and measurement of the temperature difference between challenged and control footpads. Serum antibody was characterized by fluorescence polarization. In addition, body weight was followed as an indicator of gross physiological status throughout the experiments. Materials a n d Methods
Pesticides. Dinoseb, Parathion, and pentachloronitrobenzene were analytical standards ( 1 6 ) from the Environmental Protection Agency, Triangle Park, N.C. Resmethrin, piperonyl butoxide, and mixed natural pyrethrins were generously donated by FMC Corporation, Agricultural Chemicals Group, Richmond, Calif., and Aroclor 1260 was purchased from Chem Service (West Chester, Pa.). Methotrexate (Lederle Laboratories, Pearl River, N.Y.) was chosen as a typical immunosuppressant both because of its potency in this respect and also because of its wide use in tumor chemotherapy ( I 7 ) .T h a t is to say that a considerable number of people exist who have been treated with methotrexate and who may also be e'xposed to pesticides. Animals. Inbred male hamsters, Strain LHC/LAK (Lakeview Hamster Colony, Newfield, N.J.), 5 to 8 weeks old, a n d weighing about 100 g each, were used for all experiments. Preparation of Immunogen. Chicken ovalbumin ( mol, Mann Research Laboratories, New York, 5X crystallized) was treated with fluorescein isothiocyanate (1.1X mol, Isomer I, Sigma Chemical Co., St. Louis, Mo.) in 1 M carbonate buffer, pH 9.3, for 16 h a t 4 OC. Unbound fluorescein was removed on Sephadex G-25 (Pharmacia Inc., Piscataway, N.J.), and the labeled protein (FO) was found to contain 1 mol of fluorescein per mol of protein, utilizing the molar extinction coefficients for fluorescein previously determined (18) and assuming values for ovalbumin of 7.34 for E2au1'c and of 46 000 for the molecular weight. The FO solution was stored a t -20 "C until used. T o prepare FO for injection, a solution containing 20,2, or 0.4 mg/mL (in 0.15 M NaCl) was homogenized with an equal volume of Complete Freund's Adjuvant (CFA) (Difco Laboratories, Detroit, Mich.) by means of two syringes connected together by a plastic three-way stopcock (Type K-75, Pharmaseal, Toa Alta, Puerto Rico).
0013-936X/80/0914-0204$01.00/0
@
1980 American Chemical Society
Pesticide Administration. Each animal received one dose of pesticide or other compound equal to one-half of the LDso ( 1 6 )dissolved in 1 m L of corn oil. T h e pesticide solution was administered as a bolus by intragastric feeding tube 24 h after the first injection of FO. T h e animals were fasted during this 24-h period with water ad libitum. Immunization. For all immunizations, 0.2 mL of a mixture consisting of equal volumes of FO and CFA was injected subcutaneously into each flank. For primary immunizations, FO a t either 10 or 1 mg/mL (final concentration in homogenized mixture) was used, resulting in a dose of eit her 4 or 0.4 mg per hamster. Booster immunizations consisted of 0.2 mL of FO-CFA mixture containing 200 pg of FO/mL into each flank. T h e primary immunization was given 24 h after the pesticide; boosters were administered a t 7-day intervals after the primary immunization. Serum Preparation. All blood was drawn by cardiac puncture and allowed to clot for 2 h a t room temperature and overnight a t 4 "C; the serum was drawn off after centrifugation. A preimmunization bleeding was obtained before pesticide administration and thereafter before each booster immunization. Immunoglobulin Preparation. Serum pooled from several individual bleedings was fractionated by ammonium sulfate precipitation to give a n immunoglobulin preparation substantially free of serum albumin (which if present in sufficient concentration would interfere in t h e titrations by binding fluorescein nonspecifically). T o one volume of serum, 0.58 vol of saturated ammonium sulfate (adjusted t o p H 8.1 to 8.2 by the addition of concentrated NHdOH) was added rapidly while being mixed, a t room temperature. T h e precipitate was immediately centrifuged a t 16 000g for 30 min at 20 "C. After we removed the supernatant fluid, the precipitate was dissolved in 1 serum volume of 0.15 M NaCl containing 0.001 M NaN:j. Buffer containing 0.15 M NaCI, 0.01 M KZHP04,0.005 M KH2P04, 0.001 M NaN:j, and 0.1 mg/mL rabbit y-globulin (Schwarz/Mann, Orangeburg, N.Y.) was used for the fluorescence polarization titrations, and Hepes-buffered Hanks (Flow Laboratories, Rockville, Md.) was used for footpad injections. Quantification of the Cellular Immune Response. T h e cellular immune response plays a major role in delayed hypersensitivity ( D H ) reactions, tissue transplantation rejection, and tumor immunity. Various assays such as macrophage migration inhibition, ,jlCr release from sensitized target cells, and colony inhibition are widely used as labor-saving in vitro correlates of these phenomena. However, direct measurements of the cellular immunocompetence of the whole animal are highly desirable in assessing the potential importance of pesticide effects on humans and animals. Quantification of D H reactions offers the most simple route toward assays of this type, and these have historically usually involved visual observation of t h e site and appearance of swelling and redness in an immunized animal a t the site of challenge with antigen. T h e quantitative radioisotopic footpad assay (19) represents a major improvement over previous methods for quantifying D H reactions, but does not allow continued observations on t h e same animal over a period of time. In the present study we have used methodology which seems to combine many of t h e best aspects of classical measurements with t he improvements of the Boone radiometric footpad assay. Test animals were challenged with 0.1 m L of FO (400 pg/mL in Hepes-buffered Hanks) in one footpad and with the buffer alone in the contralateral footpad. Twentyfour hours later the response was quantified in two ways. I11 the first the immune response was graded subjectively on the basis of color and swelling on a scale of 1through 4 and plotted
as a n average difference between experimental and control feet for all animals in the group (usually 5 animals). I n the second method (thermometric footpad assay), thermocouples were attached with adhesive to the test and control feet and the difference in footpad temperature was measured electronically. In some cases t h e footpads were dissected after temperature measurement in order to determine the degree of correlation between temperature and histological findings. Quantification of the Humoral Immune Response. I n order to adequately assess the humoral immune response. several parameters of circulating antibody must be measured. These include serum concentration, binding af'finity. and antibody heterogeneity. These quantities can he estimated by using radiolabeled antigen in conjunction with some separation method such as precipitation or solid-phase reaction. However, all of these methods may perturb the equilihrium state heing measured. Equilibrium dialysis circumvents this difficulty, h i t the method is slow and laborious. Fluorescence polarization (-30-22) combines many of the best features of existing methods in t h a t measurements are made rapidly without separating t h e bound and free forms of antigen or hapten. Immunoglobulin (Ig) equivalent to 10, 30.or 100 p L of serum was diluted in a fluorescence cuvette to :1 mL with buffer and the blank fluorescence measured by a fluorescence polarimeter ( 2 3 ) .Fluorescein ( 3 . 10, or 80 p L of lo-" M ) was added to the diluted Ig, and the solution was mixed with a Pasteur pipet. After 30 min a t room temperature, the fluorescence intensities and polarizations were measured. Fluorescence parameters for free. unbound fluorescein were measured in the presence of normal Ig only. T h e data were treated as described in the Appendix t o give the serum antibody site concentration. the antibody-hapten h i d i n g affinity. a n d the antibody heterogeneity index. Alternatively. the polarizations themselves can be viewed as a kind of titer assessing the immune response.
R 6's u 1t s The effect of pesticides on the cellular immune response was measured by two methods, the first being a visual evaluation of the intensity of inflammation and swelling of an antigenchallenged footpad as compared to the contralateral pad treated with buffer alone. Figure 1 shows that h y this measurement Dinoseb and Parathion markedly depress the response, while Resmethrin gives a large stimulation appearing very early after primary immunization. h,Iethotrexate and PCNB appear also to stimulate cellular immunit>-.hut only late in the immune response. Aroclor, piperonyl hutoxide. and mixed pyrethrins have little if any effect. Histologic exaniination of the footpad reaction revealed features characteristic of a delayed hypersensitivity response. T h e initial cellular infiltrate was mild and predominately granulocytic. This reached a maximum around 24 h. at which time many polymorphonuclear leucocytes were present in the deep dermal connective tissue. A mononuclear cell infiltrate hegan appearing after 12 h and increased dramatically hy 24 h after challenge. By 38 h after challenge. the mononuclear cells (plasma cells, macrophages, and lymphocytes) were densely infiltrating the dermal connective tissue to marked perivascular accumulations. T h e cellular infiltrate at this time was over 80% mononuclear. Mitotic activity was evident. and m (1 r p h o log i c changes, in c1u d in g t ran s fc r m a t i ( 111 t (1 ret i cu 1a r cells, were noted. T h e control footpads did demonstrate a mild leucocytic infiltration a t 12 h. hut a predominantly moncmuclear cell infiltrate did not subsequently develop. T h e second method for evaluating cellular immunity involved a differential temperature measurement hetween the anti~en-challeiigedand control footpads. T h e results of this Volume 14, Number 2, February 1980 205
I
Parathion ( L 1
I
Control ( L )
1
PCNB ( H )
I
Control (HI
I Piperonyl
2
"I
A
2
"1
A
Butoxide ( L )
I
2
a5 0
0
0
0 2
0
0 IO
20 30 40
IO
20
30
40
IO
Days
Figure 1. Effect of pesticides on the cellular immune response as measured by visual evaluation on a 1 to 4 scale of the inflammation and swelling of the footpad challenged with antigen as compared to the contralateral pad challenged with buffer alone. Animals were treated on day 0 with pesticide in corn oil or with corn oil alone (control) and immunized with either a low dose (L)or high dose (H) of FO as described under Materials and Methods. The curve of each pesticide must be compared with the appropriate control
thermometric footpad assay were found to correlate positively with visual evaluation of the degree of inflammation and with pathologic findings in the antigen-challenged footpad tissue. Results of the thermometric footpad assay are shown in Figure 2. Resmethrin as before shows an early, sometimes very large, stimulation, while Methotrexate, PCNB, and possibly pyrethrins give a late stimulation. Both Dinoseb and Parathion are depressive, while the other pesticides show no detectable effects. T h e effect of size of the first dose of immunogen is obviously quite important as shown by Figures 1 and 2. T h e smaller dose (0.4 mg per hamster) resulted in a much larger response than did the larger dose (4 mg). Pesticide effects on the humoral response were assessed by fluorescence polarization measurements after adding fluorescein to an Ig preparation from serum. T h e polarization itself can be thought of as a titer dependent upon both antibody concentration and antibody binding affinity. The polarization titers of Figure 3 show depression by Dinoseb and Parathion and not much effect of the other pesticides. T h e effect of size of immunizing dose is again evident, the smaller dose giving the larger response. All examined preimmunization bleedings gave only background polarization values (0.027, the same as in buffer alone). As mentioned above, Polarization is dependent upon both the antibody concentration and the binding affinity. These two variables together with a third variable, the heterogeneity index, can be segregated by an analysis of the complete titration curve. T h e results of these computations are shown in Figures 4 and 5 . In Figure 4 the quantity F h , m a x , which is equal to the molar concentration of antibody combining sites in a 300-fold dilution of serum, is shown for different pesticides during progression of the immune response. A marked depression can be seen for Dinoseb and Parathion, with only minor differences from controls for the other pesticides. In Figure 5 the largest effect on the binding affinity as measured 206
Environmental Science &. Technology
20
30
40
IO Days
20
30
40
Figure 2. Effect of pesticides on the cellular immune response as measured by average values of temperature differences (AT)between the footpad challenged with antigen and the contralateral pad challenged with buffer alone. Animals were treated with pesticide in corn oil or with corn oil alone (control) and immunizedwith either a low dose (L) or high dose (H) of FO as described under Materials and Methods
-
"I
i Methotrexote (H)
20
1
I
Pvrethrins (H)
40
20
40
Doys Figure 3. Effect of pesticides on the humoral immune response as measured by polarization titers. The polarization of fluorescence was measured 30 min after adding iop9 M fluorescein to a solution of lg (equivalent to 100 p L of serum) in 3 mL of buffer. The polarization is a function of both antibody concentration and antibody binding affinity and is synbatic with both of these variables. Animals were treated on day zero with pesticide in corn oil or with corn oil alone (control) and immunized with either a low dose (L) or high dose (H) of FO as described under Materials and Methods
by the average association constant seems to be produced by varying the size of immunizing dose. All values of KO lay between lo8 and lo9 L mol-', which is a rather small variation.
I
1
Control ( L )
I
Parathion ( L )
5L-4 Aroclor ( L )
1
109 M
Figure 4. Effect of pesticides on the humoral immune response as measured by the concentration of serum antibody against fluorescein as a function of time after primary immunization. The quantity is the molar concentration of antibody combining sites specific for fluorescein present in a 300-fold dilution of serum. Values of Fb,mx were computed from fluorescence polarization measurements; cf. the Appendix
1
1
m
'
0
Figure 6. Typical appearance of fluorescence polarization tirration curves showing polarization, p, as a function of M, the final molar concentration of fluorescein in the titration cuvette. The data points are for piperonyl butoxide at 36 days, and the smooth curves are theoretical for Fb,,,, = 6.74 X lo-' M, KO = 5.13 X lo8 M-', a = 0.77, p, = 0.0272, pb = 0.45, and Q/Q, = 6.6. See the Appendix for a discussion of these quantities and their determination
I
Parathion ( L )
I Control (H)
10
5
Days
isI
Pi0 Butox ( L )
pa
Pyrethrins ( H I
Methotrexate ( H I
1
Pyrethrins ( H I
m m
Dinoseb ( L )
~
Resrnethrin (L)
w 0
20
20
40
/
I' 40
Days
2
4
6 Weeks
2
4
6
Figure 5. Effect of pesticides on the humoral immune response as measured by the average association constant (KO)of the serum antibody present against fluorescein at different times after primary immunization. The values of KO were computed from fluorescence polarization measurements; cf. the Appendix
Figure 7. Effect of pesticides on body weight as a function of time after pesticide administrationexpressed as a percentage change of the initial weight. Animals were treated with pesticide in corn oil as indicated or with corn oil alone (control) and immunized with either a low dose (L) or high dose (H) of FO as described under Materials and Methods
However, the ohserved depressions of K Oby Dinoseb, PCNB, and pyrethrins may be significant. T h e typical appearance of fluorescence polarization titration curves can be seen in Figure 6 for piperonyl butoxide a t 36 days. T h e agreement between experiment a n d theory is very close (x2for these data is 1.19).For the other data shown
in Figures 4 and 5 , the values of x2 varied from 0.23 t o 2.2. A third variable, a , the heterogeneity index, was also obtained from the titration data. T h e values ranged from 0.7 to 1 a n d are shown in Table I with results of all the computations. T h e general physiological state of the animals was monitored by the change in body weight during the experiments Volume 14, Number 2, February 1980
207
Table 1. Effects of Pesticides on the Immune Response as Shown by Parameters Pertaining to Serum Antibody a dose Of
pesticide
immunogen
days
Pb
QdQb
control control control control control Methotrexate Methotrexate Aroclor Aroclor Dinoseb Parathion Parathion PCNB PCNB pip. butoxide pip. butoxide pyrethrins pyrethrins Resmethrin Resmethrin
L L L H H H H L
31 46 52 46 52 46 52 29 36 49 42 49 46 52 29 36 46 52 29 36
0.48 0.45 0.42 0.40 0.40 0.36 0.42 0.43 0.44 0.42 0.45 0.45 0.42 0.40 0.45 0.45 0.40 0.42 0.46 0.47
7.1
L L L L H H
L L H H L
L
5.0 3.1 4.0 4.0 6.2 5.9 6.4 4.5
5.0 5.8 4.4 5.8 8.6 6.4 6.6 5.9 4.7 5.8 5.0
a
10gFb,rnax
X2
0.75 0.77 0.77 1 0.85 0.85 0.77 0.70 0.72 0.93 1.o 0.77 0.87 0.83 0.76 0.77 0.91 0.89 0.80 0.83
0.90 4.5 6.5 0.86 4.7 9.9 5.1 5.4 6.3 1.11 0.41 1.3 4.1 4.0 5.0 6.7 2.0 3.5 4.7 9.0
1.38 1.49 1.78 1.06 0.37 0.47 1.86 1.11 0.76 1.02 1.41 0.69 2.40 0.42 1.23 1.19 0.35 0.76 1.17 1.24
lO-’Kg
5.0 0.9 3.8 2.0 0.20 0.77 0.76 3.1 3.3 2.3 6.6 1.4 1.oo 1.15 2.2 5.1 0.67 0.35 5.5 10.1
L (low) = 0.4 mg of immunogen/hamster H (high) = 4 mg of irnmunogen/hamster.pb, the polarization of bound fluorescein. Of/&, the fluorescence ratio of free to bound. KO,the association constant, L mol-’. a, the heterogeneity index. the antibody site concentration in 300X diluted serum. = 1 O 0 ~ ( p o b d a
- pca,cd)P. statistical measure of fit between observed and calculated polarization.
(Figure 7). Both Dinoseb and Parathion show a prolonged effect in depressing growth, while Methotrexate and Arochlor give transient depressions. The high dose of immunogen itself shows a marked depression when compared to t h e low dose control. Discussion The experimental data obtained show that a single dose of orally administered pesticide may exert large, long-lasting effects on the immune response. T h e effects observed may be either stimulation or depression and may be directed selectively toward either the cellular or humoral immune response depending upon the pesticide. In this study these effects were monitored by measurement of several parameters related to various aspects of the immune response to a single well-defined immunogen, fluoresceinlabeled ovalbumin (FO). The humoral response measured was directed against fluorescein itseIf, while the cellular response was that directed against the entire immunogen. T h e effects observed as summarized in Table I1 show that t h e most significant findings are a marked depression of both the cellular and humoral immune response by Dinoseb and Parathion and an early and sometimes very pronounced stimulation of the cellular response by Resmethrin. T h e latter effect is of considerable interest and may provide an important clue toward designing potent stimulators of the cellular immune response. Such materials could be of great value in treating bacterial, viral, and certain neoplastic diseases. The actions of Parathion and Dinoseb are remarkably long-lasting and the depression of the immune response which they evoke could lower resistance to a variety of infectious diseases. Inhibitory effects of Dinoseb, Parathion, PCNB, and pyrethrins on the humoral response are seen to reside chiefly in the amount of antibody produced and not upon its binding affinity. Probably the type of antibody is unchanged by these pesticides but the amount is decreased. T h e results of this study provide important leads which should be followed up. First, the immunological effects of 208
Environmental Science & Technology
many other pesticides and organics generally should be investigated and, secondly, the effects should be studied not only in the whole animal but also a t the level of T-cell and B-cell activation and macrophage function. Appendix In this Appendix a brief review of fluorescence polarization is given together with the necessary equations for interpreting polarization data and procedures for computing several derived parameters. The essential feature of applying fluorescence polarization to antibody hapten binding consists in observing the degree of polarization and intensity of the fluorescent light when measured quantities of the hapten and antibody are allowed to interact (20-22). From measurements of polarization, the final, derived parameters that can be obtained are (1)binding site concentration, (21 an average association constant, and (3) an index of the heterogeneity of the binding sites. T h e general type of reaction assumed is that in which a fluorescent ligand, -7, binds to a receptor. A, to reversibly form a complex, .7 I{:
3
+ i{ z3/3
(la)
Symbols used are: a , heterogeneity index; b, subscript indicating “bound”; F , molar concentration of 3 ; f, subscript indicating “free”; Fb,max,the maximum value of F h , taken to be equal to the total molar concentration of receptor sites; M, the total molar concentration of S in both free and bound forms (AM in computer program); p , the polarization of the excess fluorescence, Le., p = (Au - Ah)/(Au A h ) , where Au and Ah are the intensities in arbitrary units of the components in the excess fluorescence (above that of the blank) polarized in t h e \rertical and horizontal directions, respectively: Q, molar fluorescence of a mixture of free and bound forms of 3 as they exist in a solution under observation, i.e., Q = (Au
+
+
Ah)/M. Fluorescence polarization and intensity measurements provide a measure of the bound/free ratio:
fable II. Summary of Effects of Pesticides on the Immune Response a cellular pesticide
thermometric footpad
Methotrexate
+ (late)
+ (late)
Aroclor
0
Dinoseb Parathion PCNB pip. butoxide
-
0 -
pyrethrins
Resmethrin a
visual evaluation
-
+ (late)
+ (late)
0 0
0
+ (early)
f (late) (early)
+
- polarization
humoral antibody concn
liter
0 0
0 0
binding afflnily
0 0
-
-
0 to 0
0 0 0
0
0
0 0
0 Oto-
0
0
0 to
-
+
-,depression. f ,stimulation. 0, iittie or no effect.
estimates have been made for the five parameters, the program proceeds to improve these estimates in a n iterative fashion until a stopping criterion specified by the user is met. T h e measure of goodness to fit to Equation 7a is a modified X-square defined by:
and
In order to utilize these equations, the constants Qt, 4t1,p t , and pi, must be determined for a particular system under study. No problem is posed in finding Qt and pi, since these come directly from a measurement on the labeled component alone. T h e determination of 41, and pt,, however, implies measurements on a state in which the fluorescent labeled material is completely bound to its complementary partner. Since complete binding cannot be realized physically. an extrapolation is involved. If equilibrium values of p plotted against M are extrapolated to M = 0, p approaches a limit, p ' . Values of p' for different antibody concentrations plotted against ( p ' - P I ) divided by antibody concentration in any arbitrary convenient units give P I ,as the intercept of a straight line, for classical mass law ( 2 4 ) :
This procedure makes it unnecessary to know absolute values of KF~,,,,,i~x beforehand. A similar relationship facilitates the determination uf 41,:
4' = 41, + 41- Q' ~
(3a)
KFII.",,, In the algorithm given below, the measured values of M , p , Q, p f ,Q f , and relative antibody concentrations (AB) with or without tentative estimates of p h and Q h are used in an iterative computation to derive the final best values ofpt,, 4t,, KO, a,and F h , m a x . These computations are based upon achieving a X-square fit of t h e data to the Sips equation (24,%), which defines Ff as:
Substituting M = give:
FI, + F l into Equation 2a and rearranging
Inspection of Equations 6a and 'ia shows that there are five unknowns, F h , m a x , a , K O ,P h , and 4 h , to be evaluated from measured individual values of p , M , and 4, viz., pi, Mi,and Qi, and from the measured values of p f and 4f. In the procedure given below the user may either specify initial estimates for these unknowns along with measured molarity and polarization data, or the user can allow the program POLAR to make the initial estimates. Once initial
CHISQ =
(Pi- Pcalcd,i)' 0.01
where p is t h e measured polarization and Pc&d is the value computed from Equation 7a given the current estimates for the five parameters. T h e iterative improvement performed by POLAR consists of repeatedly moving away from t h e current best estimate of each of the parameters either by a user specified value (if the user provided the initial estimates) or by 10°C of the current estimate (if POLAR computed t h e initial estimates), computing the X-square value t h a t results, and keeping track of the parameters t h a t give t h e smallest X-square value. As t h e iteration proceeds, t h e amount t h a t is being added or subtracted from the current best estimate is halved as the value of the parameter nears an opl imal value. T h e iteration continues until all the parameters satisfy t h e following test, where RANGE is a value entered by the user a t the onset: increment current value
< RANGE for all five parameters
(9a)
A frequently occurring computation in this search for optimal parameters is the value of Ft,. This is accomplished by a root-finding routine ZEROIN, which will find t h e zero of a function, given a n initial interval in which that root must lie (26').Equation 6a for F f , is written as follows t o avoid t h e singularity present when F,,, approaches Fll max:
+ Ft,) - AM + F,, = 0 using the fact that molarity AM M = Ft,, + FiL. iK,lFt,)('(Ftlmdy - AM
(loa)
T h e program POLAR operates in two modes depending on the amount of data the user can supply. Option 1. T h e program reads in pf,Qf, N (= number of antibody concentrations), and RANGE. Then, for each antibody concentration, t h e following data are read: AB = antibody concentration (scaled so smallest is l ) ,N P = number of molarities a t which p and Q have been measured, followed by N P sets o f AM M = molarity; p = polarization; Q = molar fluorescence. Using these data, the program proceeds to: Extrapolate each set of p and Q data t o a zero molarity value (called p' and 4') using a subroutine which performs Lagrange interpolation (27). Therefore, for each antibody, there is p' = p ( 0 )and Q' = Q(O). Obtain a least-squares fit to a straight line y = a x b. Use a subroutine that assembles and solves the normal equations, which are:
+
Volume 14, Number 2, February 1980 209
to obtain initial estimates of Pb and Qb, which are defined as the intercepts of:
and
Iteratively find initial estimates for Fb,max, K O ,and a (renamed in the program: FBMAX, OK, A). Use the X-square fit to improve these initial estimates. Option 2. T h e program reads in pf, Qf, N ( = number of antibody concentrations), and RANGE. Then for each antibody the following data are read: AB = antibody (scaled), N P = number of molarities at which p has been measured followed by N P sets of AM M = molarity and p = polarization. The program then reads in initial estimates provided by the user for the five parameters: FBMAX, QF/QB (only ratio is important), A, OK, and PB, followed by five values specifying how much the current values of the parameters are to be varied in optimizing the X-square fit. If the parameter is not to be changed by the program, then a zero should be entered for the corresponding variance. Otherwise a good value to use is 10%of the original estimate. T h e program proceeds then to improve these initial estimates using the X-square criterion.
Literature Cited (1) Ercegovich, C. D.,Fed. Proc., Fed. Am. SOC. Riol., 32(9), 2010-6
(1973). (2) Rosival, L., Barlogova, S., Grunt, J., Gig. Tr. Prof. Zabol., ( 6 ) ,53-5 (1974) (Russian). (3) Latimer, J. W., Siegel, H. S., Poult. Sci., 53(3), 1078-83 (1974). (4) Shtenberg, A. I., Khovaeva, L., Zavarzin, M. V., Vopr. Pitan., (4), 35-42 (1974) (Russian). (5) Wassermann, M., Wassermann, D., Kedar, E., Djavaherian. M., Bull. Enciron. Contam. Toxicoi., 6(5), 426-35 (1951). (6)Wassermann, M., Wassermann, D., in “Fate of Pesticides in
Environment”, Tahori, A. S.. Ed., Gordon and Breach, New York, 1972. p p 521-9. ( 7 ) LVassermann, M., Wassermann, D., Kedar, E., Djavaherian, M., Cucos, S., Ventura, S., Hull. Enciron. Contam. Toricol., l O ( 1 ) . 42-50 (1973). (8)Wassermann, M., Wassermann, D., Kedar, E., Djavaherian, M., Cucos, S., Huli. Enciron. Contam. Toxicol., 8(3),177-85 (1972). ( 9 ) Haugen, J., Acta Patho/. h!icrobio/. Scand., Ser. R, 79, 219-25 (1971). (10) \.’os, .J. G., Van Driel-Grootenhuis, L., Sci. Total Enciron., 1(3), 289-302 (1972). (11) Rolkhovityanova, V. M., Aleksevich, Y., Vrach. Delo, 8, 116-7 (1968) (Russian). (12) Aripdzhanov, T. M., Gig. Sanit., ( 5 ) , 101-2 (1973) (Russian). (13) Aripdzhanov, T,M., Gig.Sanit., ( 7 ) ,39-42 (1973) (Russian). (14) Glick. B., Poult. Sci., 53, 1476-85 (July 1974). (15) Street, ,J. C., Sharma, R. P., Toxicol. A p p l . Pharmacol., 32(3), 587-602 (1975). (16) Thompson, J. F., Ed., “Analytical Reference Standards and Supplemental Data for Pesticides and Other Organic Compounds”, Publication No. EPA-600/9-76-012, Environmental Protection Agency. Technical Publications Branch, Office of Administration, Research Triangle Park, N.C., 1976. (17) Bertino, J . R., Cancer R e s . , 23, 1286-1306 (1963). (18) Dandliker, W. B., Alonso, R., Immunochemistrj, 4, 191-6 (19ti7). (19) Paranjpe, M., Boone, C., J . N u t / . Cancer Inst., 48,563 (1972). ( 2 0 ) Dandliker, W. B., in “Methods in Immunology and Immunochemistry”, Vol. 3, Williams, C. A,, Chase, M. IV., Eds., Academic Press, New York, 1971, pp 435-462. (21) Dandliker, W.R., in “Immunochemistry of Proteins”, Vol. 1, Atassi, M. Z., Ed., Plenum Press, New York, 1977, pp 23-61. (22) Dandliker, W. B., Dandliker, J., Levison, S. A,, Kelly, R. J., Hicks, A. N., White, J. U., Methods Enzymol., 48F, 380-415 (1978). (2:3) Kelly, R. J . , Dandliker. W.B., \t’iIliamson; D. E., Anal. Chem., 48,846-56 (1976). (24) Dandliker, W.B., Schapiro, H. C., Meduski, J. W,, .4lonso, R., Feigen, G . A,, Hamrick, J . R., Immunochemistry, 1(3), 165-91 (1964). (E?) Xisonoff, A., Pressman, D., J . Irnmunol., 80,417-28 (1958). ( 2 6 ) Forsythe, G. E., Malcolm, M. A,. Moler, C. B., “Computer Methods for Mathematical ComDutation”. Prentice-Hall. Enelewood Cliffs, N.J., 1977, pp 164-6 (27j Carnahan, B.. Luther. H. A., LVilkes, J . O., ”Applied Numerical .. hlethods”. LVViley, New York, 1969, p 31. I
Receiced for recieu. March 16, 1979. Accepted Nocember 15, 1979. Supported b j , Grant No. R803885 from the L’.S. En~,ironmenta/ F’rotPc,tion Agencj,, b), Grant .Vo. GR-31611 from the National Science Foundation, and by Training Grant N o . 4 M 07097from the Nationai Instilutr>sof Health.
Effectiveness of a Barrier Net in Reducing White Perch (Morone americana) and Striped Bass (Morone saxatilis) Impingement Stephen J. Edwards” Lawler Matusky & Skelly Engineers, One Blue Hill Plaza, Pearl River, N.Y. 10965
Jay 6. Hutchison Jr. Orange and Rockland Utilities, Inc., One Blue Hill Plaza, Pearl River, N.Y. 10965
A net was deployed as a barrier in front of a water intake located on an embayment of the Hudson River during periods of increased young-of-the-year and yearling ( 5 to 10 cm total length) white perch (Morone americana) and striped bass (Morone saxatilis) impingement. Multifilament nylon nets of two mesh sizes (1.27 and 0.95 cm) were investigated with no appreciable difference noted in clogging or debris accu-
210
Environmental Science & Technology
mulation. Several methods were used to measure the effectiveness of the barrier net, including comparisons of impingement records and release and recapture of tagged fish. Based on tag recoveries, the 0.95-cm mesh net was estimated to have reduced impingement by as much as 90%. The barrier net proved to be an easily deployed device, effective in reducing impingement. 0013-936X/80/0914-0210$01.00/0 @ 1980 American Chemical Society
