Original Articles - IOVS

Original Articles

Local antibody formation within the eye: a study of immunoglobulin class and antibody specificity Kohkichi Shimada* and Arthur M. Silverstein

Primary immunogenic uveitis was induced in the rabbit eye with a single injection of antigen into the vitreous, and secondary booster uveitis responses were induced two months later by intravenous administration of the same antigen. The distribution of immunoglobulin classes and the specificity of the antibodies produced were assessed early and late in the primary response and early and late in the secondary response, and were compared with the analogous responses in the spleen and regional lymph nodes. At each of these stages of intraocular antibody response, IgG formation was higher and IgM formation lower than that seen in organized lymphoid tissues, while the proportion of IgA-forming cells was similar to the low levels usually found in the spleen. A significant proportion of IgA-forming cells was found in the perilimbal conjunctiva, and even greater levels in the lacrimal glands. At each stage of the response, the proportion of immxinoglobulin-forming cells making antibody specific for the inciting ovalbumin antigen was surprisingly low, reaching only seven per cent during the late primary reaction and 18 per cent during the late secondary reaction. Key words: uveitis, antibody formation, immunoglobulin class, immunogenic uveitis, antibody specificity.

-Lhe introduction of a bland antigen into the vitreous of the rabbit eye elicits a nongranulomatous uveitis accompanied by the

local production within the eye of significant amounts of antibody.1'3 Even after resolution of this immunogenic inflammation, the eye is left with long-standing immunologic memory of the antigen involved, such that re-exposure to specific antigen at a later date will result in recall of the uveal inflammation and a booster antibody response within the eye. The intraocular responses to antigenic challenge are quite similar to the sequence of events that take place in organized lymphoid tissues undergoing primary and secondary antibody responses to antigen, leading to the suggestion4 that the uveal tissues of the eye may assume the functions of a regional lymph node under certain circumstances.

From The Wilmer Institute, Johns Hopkins University School of Medicine, Baltimore, Md. Supported in part by United States Public Health Service Research Grant EY 00279, an unrestricted gift from Alcon Laboratories, Inc., and an Independent Order of Odd Fellows Research Professorship. Submitted for publication Nov. 7, 1974. Reprint requests: Dr. A. M. Silverstein, The Wilmer Institute, Johns Hopkins Hospital, Baltimore, Md. 21205. "Present address: Department of Ophthalmology, Jichi Medical School, Tochigi-ken, Japan.

573

Investigative Ophthalmology August 1975

574 Shimada and Silverstein

This concept was supported by the demonstration that bits of lymph node implanted into the anterior chamber of the rabbit eye would maintain their normal lymphoid and antibody-formation functions in this ectopic location/' While the antigenic specificity of an antibody depends upon the primary amino acid sequence at the N-terminal end of its heavy and light chains, these antibody molecules fall into different immunoglobulin classes depending upon the amino acid sequence of the C-terminal end of their heavy chains.0 It has been shown7's that there exists in organized lymphoid tissue a fairly characteristic distribution of IgG, IgA, and IgM immunoglobulin classes (characterized, respectively, by the y, a, and fA. heavy polypeptide chains). Since the distribution and proportion of the various immunoglobulin classes within the eye during immunogenic uveitis have not previously been reported, the present studies of the relationship of the several immunoglobulin classes within the uveal tract to those in the regional lymph nodes and spleen were designed as a further test of the hypothesis that ocular tissues may assume the functions of organized lymphoid tissue. Data are also presented on an analogous comparison of the specificity of the antibodies produced following primary and booster stimulation of local antibody formation within the eye. Materials and methods Antigens. Ovalbumin (OA) and bovine gamma-globulin (BGG) were dissolved in saline at a concentration of 20 mg. per milliliter, sterilized by passage through a 0.22 /t Millipore filter, and stored at -20° C. without preservative. Animals. New Zealand albino rabbits weighing 5 to 7 pounds were used. Before use, all eyes were examined with the slit lamp and ophthalmoscope for the presence of pre-existing diseases. Antigen injections. Intravitreal injections were performed after topical anesthesia with 0.5 per cent proparacaine HC1. The eye was fixed by gently grasping the superior rectus muscle at the insertion with forceps, and the eye was rotated downward. A sharp 27-gauge needle was then inserted at a slight angle backward 2 mm. behind the corneoscleral junction, avoiding the

lens. The antigen solution (0.1 ml. containing 2 mg.) was injected into the center of the vitreous. The secondary challenge was performed by the intravenous injection of 1 ml. OA solution (20 mg.) about two months after the initial intraocular injection. Control injections. Studies of the primary antibody response within the eye were performed within the same animal by injection of 0.1 ml. of sterile saline into the contralateral eye. To control the study of the secondary booster response, the fellow eye was injected intravitreally with 0.1 ml. of BGG at the time of initial sensitization of the test eye with OA. Fluorescent antibody reagents. The preparation and fluorochrome labeling of anti-y, anti-«, and anti-M rabbit heavy chains has been reported.7"1 u We are indebted to Dr. John J. Cebra of the Biology Department, Johns Hopkins University, for furnishing many of the y, «, and /i chain reagents, and for his kind guidance in other preparative procedures. Anti-OA was prepared by repeated immunization of rabbits with OA in adjuvant, and its globulin fraction was conjugated with fluorescein isothiocyanate according to Wood, Thompson, and Goldstein.11 Tissue preparation and staining of specimens with reagents. Rabbits were killed by intravenous air injection. The eyes, lacrimal glands, preauricular lymph nodes, and spleens were removed. The tissues were appropriately dissected, and frozen on dry ice. Serial tissue sections (4 fi) were cut in a cryostat, using two blocks from each tissue examined. Neighboring sections were stained individually with each reagent for various rabbit heavy chains by the direct method, and with anti-OA by the indirect method. Stained specimens were examined under a Zeiss ultraviolet microscope. Positively stained cells were counted for each preparation, and the relative percentage of cells stained by each reagent was calculated. Examination of multiple sections with each reagent assured that a representative sample of the cell population within a given tissue was obtained, so that the values obtained with all reagents could be combined to yield valid percentages. No attempt was made, however, to determine the distribution pattern of the Ig-cells within a given tissue. The data presented represent the averages of the two animals in each group-

Results

The experiments were designed to assess the proportion of y, a, and /x heavy chains, as well as the proportion of cells forming antibody specific for the inciting antigen at four different stages in the anti-

Volume 14 Number 8

Local antibody formation 575

Table I. Distribution of heavy Ig polypeptide chains and anti-ovalbumin antibody among lymphoid cells of rabbits six days after a single injection of ovalbumin into right eyes and saline into left eyes Per cent heavy Ig chains* Cells counted

Anti-OAf

y

a

•ft

Tissue Right side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Left side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Spleen:

294 218 31 2 338 251

91 89 94 50 46 59

1 2 0 0 24 2

8 9 6 30 39

2

100 100 100 0 100 45 68

0 0 0 0 0 1 7

0 0 0 0 0 54 25

10 1 0 1 469 1,042

"The values represent the percentage of all 7 fThe values represent the percentage of all y

50

0 1 0 0 0 0

0 0 0 0 0 0 0

fi cells containing the respective heavy chains, /t-containing cells which stain for specific anti-OA antibody.

body response associated with the production of immunogenic uveitis: early in the primary response, late in the primary response, at the start of the secondary booster response, and late in the resolving phase of the secondary response. The primary response. Four animals were employed. OA solution (0.1 ml. containing 2 mg.) was injected into each right eye, while the control left eyes all received the same volume of sterile saline. The initial reaction to the trauma of injection had completely subsided within two days in all eyes. Three of the four right eyes developed uveitis on the sixth day after inoculation, and the other on the seventh day. Two animals were killed on the day of onset of uveitis for study of the early primary reaction at which time they showed a 1+ to 2+ aqueous flare and cells, with mild perilimbal injection and iris hyperemia. Uveitis continued to increase in severity in the other two animals, so that some four days after onset of ocular inflammation, the right eyes exhibited a fullblown uveitis with 4+ aqueous flare and cells, a severe perilimbal injection, and iris hyperemia. Following this, clinical uveitis slowly resolved so that after a further seven days the eyes were judged to be essentially

normal clinically. At this point (11 days after onset of uveitis), these rabbits were killed for study of the late primary reaction. In no instance did the control left eye of any animal show a clinical reaction to the saline injection. The early primary response. On the day of onset of uveitis, immunoglobulin-containing cells (Ig-cells) were demonstrable in large numbers in both right eyes, particularly in the stroma of the anterior uveal tract and in the perilimbal conjunctiva (Table I). These cells were primarily localized in perivascular cuffs (Fig. 1), but did not present the same picture of tight Ig-cell clusters as seen in the spleen and lymph nodes. Fluorescing Ig-cells could scarcely be found in the cornea except in the neighborhood of the limbus, and few cells were present in the posterior segment of the eye. As may also be seen from Table I, very few Ig-cells could be found within the uveal tract or conjunctival tissues of the saline-injected eyes. Almost 90 per cent of the Ig-cells seen in the inflamed eye early during the course of uveitis contained y chain, with /x chaincontaining cells present only in relatively small numbers (less than 10 per cent in all parts of the uveal tract). Among ocular



Fig. 1. The ciliary body of an eye injected with ovalbumin, taken six days later during the early primary response, and stained with fluoresceinlabeled anti-y heavy chain. Fluorescing cells were located principally around blood vessels, with a few IgG-forming cells scattered throughout the connective tissue.

Fig. 2. The ciliary body of an inflamed eye during the late stage of the primary response to intravitreal ovalbumin, stained with fluorescein-labeled anti-y heavy chain. Many fluorescing cells were present in the stroma and infiltrating the epithelial layer.

tissues, only the perilimbal conjunctiva showed a higher proportion of /^-containing cells. Cells within the eye containing a chain constituted only one to two per cent of the total, but again this class of immunoglobulins was produced very prominently in conjunctival tissues, accounting for about one-fourth of all Igcells. In the saline-injected control eyes, very few Ig-cells were found, but all of these stained positively for y chain. Compared with the values for the intraocular tissues, the proportion of ^-containing cells in the spleen and preauricular lymph nodes was higher, whereas a-containing cells were more prevalent in the spleen but not in the regional node. Despite the presence of large numbers of Ig-cells within the uveal tract, only a few cells within the iris stroma stained positively for specific antibody against OA, and at this time no cells could be found

making specific anti-OA antibody in either spleen or preauricular lymph node. The late primary reaction. Immediately following the clinical subsidence of uveitis, many Ig-cells were still demonstrable within the uveal tract of the antigen-injected right eyes, particularly among the epithelial layers (Fig. 2), with lesser numbers scattered throughout the stroma. As is seen in Table II, appreciable numbers of Ig-staining cells could now be found within the retina. At this late stage, almost all of the fluorescing cells were stained positively with the reagent for y chain (93 per cent or more), with only four to five per cent of the cells staining for /x chain and two to three per cent for a. chain. It is of some interest that most of the a- and ja-containing cells within the eye were diffusely infiltrating the stroma of the various tissues, while most of the y-containing cells were close to or infiltrating the epi-

Volume 14 Number 8


Table II. Distribution of heavy Ig polypeptide chains and anti-ovalbumin antibody among lymphoid cells of rabbits during the late phase of primary responses 17 days after single injection of ovalbumin into the right eyes and saline into the left eyes Tissue Right side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Lacrimal gland Left side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Spleen:

Per cent heavy Ig chains0

Anti-OAf

y

a

(%)

904 541 284 239 106 407 478

93 94 94 97 70 56 5

2 2 3 2 29 4 95

4 3 1 1 40 0

15 13

100 77 100 0 37 49 64

0 23 0 0 62 6

0 0 0 0 1 45

5

31

Cells counted

5 0 79 399 1,124

°The values represent the percentage of all y + a fThe values represent the percentage of all 7 +

5

10 8 3 3 1 0 0

0 0

0 0 0 0 1

n cells containing the respective heavy chains. + /j-containing cells which stain for specific anti-OA antibody.

thelial layers. At this stage in the uveitis, //.-containing cells had almost completely disappeared from the perilimbal area, while many y and a chain-containing cells were still demonstrable. At this time, the control saline-injected eyes still contained relatively few inflammatory cells, most of which stained positively for y chain, while a few a-containing cells were found in the iris and many more in a much higher proportion in the perilimbal conjunctiva of this eye. Within the lacrimal gland, a chain-containing cells made up 95 per cent of all immunoglobulinstaining cells, with y-containing cells constituting the remainder. No cells of this gland stained positively with anti-^a chain reagent. The proportions of the various heavy-chain constituents within the cells of the spleen and preauricular lymph nodes did not show any significant difference from those described for the early primary response in Table I. Table II also shows a marked increase in the number of Ig-forming cells now making antibody specific for the stimulating antigen within the test eye. As many as 10 per cent of these cells within the ciliary

body (Fig. 3) and three to eight per cent of these cells within the retina, choroid, and iris were now found to be making specific anti-OA antibody. However, no specific anti-ovalbumin-forming cells could be found within the conjunctiva of the injected eye nor anywhere within the control eye or preauricular lymph nodes. Only in the spleen was specific antibody formation found, to the extent of only one per cent of Ig-containing cells. The secondary booster response. Primary uveitis was induced in the right eye of four rabbits by intravitreal injection of 2 mg. of OA. In this instance, the control left eyes received 2 mg. of BGG intravitreally at the same time. Both right and left eyes of all rabbits developed a primary uveitis about one week later, and this was permitted to resolve. About two months after the initial intraocular injections, secondary challenge of all rabbits was performed by administration of 20 mg. of OA solution intravenously. In each animal, only the right eye developed uveitis within 24 hours after the intravenous challenge, while no inflammatory response was observed in any of the left eyes which had earlier re-

578 Shimada and Siherstein

Fig. 3. The ciliary body of an ovalbumin-injected eye taken during the late stage of the primary response and stained for anti-ovalbumin antibody formation, using the indirect method with fluorescein-labeled reagents. Evans blue counterstain was employed to allow characterization of the background cells. A few brightly fluorescing cells forming specific antibody may be seen on a background of an extensive lymphocytic inflammatory cell infiltrate.

ceived BGG. Two rabbits were killed on the day of onset of the secondary uveitis for an assessment of their antibody responses at this time. Clinical uveitis continued to develop in the right eyes of the remaining two animals, reaching a peak on the second or third day and then slowly resolving. One week after secondary challenge, ocular inflammation in the right eyes appeared to have subsided and the eyes were substantially normal clinically. The remaining two rabbits were killed at this time. The early secondary response. It is apparent from Table III that as early as 24 hours after specific intravenous booster immunization, large numbers of immunoglobulin-forming cells could be found within the tissues of the right eyes which had experienced the priming reaction two months earlier. Again, most of these Igcells were producing y chain (86 to 100 per cent), while lesser numbers were forming a chain. At this time, the cells forming IgG and IgA were distributed primarily around vessels of the uveal tract, while the smaller numbers of ^-forming cells were


Fig. 4. The perilimbal conjunctiva taken early in the primary response to intravitreal ovalbumin and stained with rhodamine-labeled anti-« heavy chain. Many orange fluorescing cells forming IgA were scattered throughout the subconjunctival tissues, with a few such cells infiltrating the epithelial layer.

more uniformly distributed throughout the tissues. The appearance and distribution of a chain-forming Ig-cells within the perilimbal conjunctiva of the right eye may be seen in Fig. 4. Once again, the perilimbal conjunctiva showed appreciably more aforming cells (30 per cent) than did the intraocular tissues, while almost all of the Ig-cells within the lacrimal gland (91 per cent) were found to be IgA-forming cells. The control left eyes in this experiment, which had experienced a BGG-specific uveitis some two months earlier, appeared not to have participated significantly in the OA-specific booster response, since far fewer inflammatory cells were found within the intraocular tissues on this side. Most of the Ig-cells within the control eye were producing y chain, and none were found to produce /x chain. The large number of a chain-producing cells in these eyes was somewhat surprising, reaching upwards of 20 per cent in the choroid and ciliary body.

Volume 14 Number 8


Table III. Distribution of heavy Ig polypeptide chains and anti-ovalbumin antibody among lymphoid cells of rabbits during the early phase of secondary responses on intravenous challenge with ovalbumin two months after the initial injection of OA into right eyes and BGG into left eyes Per cent heavy Ig chains' Tissue Right side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Lacrimal gland Left side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Spleen:

Anti-OAf

Cells counted 371 169 144 40 234 833 140

95 86 91 100

4 8 8 0

68

30

52 8

3 91

49 23 21 12 62 1,125 1,136 i.,UO

78

22 9 20 0 37 2 10 IV

91 80 100 63 41 70 IV

1 6 1 0 2 45 1 0 0 0 0 0 57 20 ZU

4 6 1 10 0 0 0 0 0 0 0 0 0 0 V

°The values alues represent thi the percentage of all y + a + fi cells containing the respective heavy chains. tThe 'alues represent the percentage of all y + a + /i-containing cells which stain for specific nnti-OA antibody.

The distribution of the several chains within the preauricular lymph nodes and the spleen appears to be typical for these tissues. Modest numbers of Ig-cells forming specific anti-OA antibody could be found scattered diffusely through the uveal tract and retina of the right eye, but were completely absent in the control left eye and in the right perilimbal conjunctiva. At this stage in the response, the formation of specific antibody within organized lymphoid tissue was so limited that only a single anti-OA-forming cell could be found in the right preauricular lymph node and the spleen among all of those Ig-cells counted. The late secondary response. During the subsiding phase of the booster response, many Ig-cells were still demonstrable within the uveal tract, particularly in the anterior segment (Table IV). Approximately 95 per cent of the fluorescing Igcells were stained with the reagent for y chain, while /x chain-containing cells were rare within these tissues. The proportion of cells with a chain within the intraocular tissues was higher during the secondary response (three to ten per cent) than during

the primary response. Of the Ig-cells within the control eye, about 90 per cent were y-containing and the remainder contained a chain, with no p, chain-containing cells found. The distribution of heavy immunoglobulin chains within the perilimbal conjunctivae and the preauricular lymph nodes and spleen was comparable to that seen in other experiments, although there seemed in this late secondary reaction to be a shift in the direction of cells containing y chain within the spleen. Table IV shows also that it was at this late stage of the booster response that the maximum numbers of specific anti-OAforming cells were found within the intraocular tissues of the test right eye, although even at this time cells making specific antibody averaged only some 18 per cent of the Ig-cells within intraocular tissues. At this time also, some 9 per cent of the Ig-cells within the perilimbal conjunctiva of the test eye were also making specific anti-OA antibody, although no specific anti-OA antibody could be found within the intraocular tissues of the control left eye and only a single cell within its perilimbal conjunctiva.



Table IV. Distribution of heavy Ig polypeptide chains and anti-ovalbumin antibody among lymphoid cells of rabbits during the late phase of secondary responses on intravenous challenge with OA two months after the initial injection of OA into right eyes and BGG into left eyes Per cent heavy Ig chains0 Tissue Right side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Lacrimal gland Left side: Ciliary body Iris Choroid Retina Perilimbal conjunctiva Preauricular lymph node Spleen:

Anti-OAf (%)

Cells counted

y

a

523 445 97 14 47 1,057 307

97 95 90 93 79 51 29

3 4 10 7 19 5 71

0 1 0 0 2 44 0

14 20 23 43 9 37 0

67 42 18 4 91 1,201 1,178

88 95 83 100 67 60 83

12 5 17 0 32 4 4

0 0 0 0 1 36 13

0 0 0 0 1 29 34

"The values represent the percentage of all 7 + a + M cells containing the respective heavy chains. fThe values represent the percentage of all y + a + /^-containing cells which stain for specific anti-OA antibody.

The widespread involvement of organized lymphoid tissues in the anti-OA response at this time is witnessed by the relatively high proportion of specific antibody-forming cells within the ipsilateral and contralateral preauricular lymph nodes and the spleen. The lymph nodes were found to contain 29 and 37 per cent of Ig-cells making specific anti-OA antibody, while the spleen showed that 34 per cent of its Ig-cells were so occupied. Discussion Following the induction of immunoo

genie uveitis stimulated by the intravitreal injection of antigen in the rabbit eye, cells forming antibodies specific for the inciting antigen appear in small numbers. The proportion of inflammatory cells devoted to specific antibody formation increases rapidly and reaches a peak just after the subsidence of clinical uveitis.2*3 The kinetics of this local ocular antibody response, and particularly the ability to induce a secondary booster antibody production within the eye in response to parenteral administration of the same antigen at a later date, suggested that ocular tissues

might be capable of immunologic responses comparable to those seen in the organized lymphoid tissues of spleen and regional lymph nodes. The proportions of immunoglobulin heavy chains of the y, a, and /A classes found in the spleen and regional lymph nodes are similar to those reported by earlier investigators.7's> 12 In the spleen, some 65 to 80 per cent of all Ig-forming cells are found to contain y chain, with 13 to 30 per cent containing ^ chain, and 4 to 10 per cent a chain. The distribution in the regional lymph nodes is somewhat different, in that appreciably more IgM formation is seen at the expense of IgG, while the number of a chain-containing IgA cells is generally less in the lymph node than in the spleen. Antibody formation within intraocular tissues is different from that seen in lymphoid tissues. At all stages during the course of both primary and booster intraocular responses, the proportion of y chain-containing IgG cells is invariably much higher, usually 90 per cent or more. At the same time, the proportion of IgM-forming cells is appreciably lower than for either

Volume 14 Number 8

the spleen or lymph node, being some eight to nine per cent during the early primary response, diminishing to three to five per cent during the late primary response, and reaching only some one to two per cent during the secondary response. At the same time, the proportion of IgAforming cells within ocular tissues starts out at a low one to two per cent during the early primary antibody response, and increases to some four to eight per cent during the secondary booster antibody response within the eye. It is a well recognized phenomenon that the stimulation by antigen of immunocyte precursors within lymphoid tissues leads to an early IgM response, followed quickly by a shift to IgG antibody formation. Since the lymphoid tissues of the host are continuously responding to a variety of different antigenic stimuli in addition to the antigens which we have introduced into the eye and intravenously, we may assume that the proportions of cells staining positively for the several immunoglobulin classes represent the summation of all of these immunologic responses. However, antibody formation within the eye differs in two significant respects from that which occurs in organized lymphoid tissue elsewhere. First, we must assume that the stimulus for the ocular response is restricted to the antigen introduced within the eye, and that there does not occur in these isolated tissues the continual stimulus and re-stimulus by other exogenous antigenic agents.1"1 Second, it seems clear13-14 that a different population of cells supports antibody formation within the eye as compared with that which participates in the lymphoid tissue response. In lymphoid tissues, all of the precursor cells are presumably present throughout, and respond in situ to the antigenic stimulus. In contrast, ocular tissues normally possess no immunocyte population,14 and all cells which participate in the inflammatory reaction and its accompanying local antibody formation immigrate into the ocular tissues through the blood.1H' 14 Thus the primary


effect of intraocular antigen occurs only after it leaks out of the eye and stimulates immunocytes systemically. Those stimulated cells then find their way into the circulation and, upon interaction with residual antigen within the eye, mediate inflammation and differentiate further into antibody formation. The preponderance of y chain-containing Ig-cells within the eye during uveitis may, therefore, merely reflect the fact that the typical IgM component of the immune response has remained within the spleen, and that IgG precursor cells circulate more readily than IgM cells. Since the primary antibody response within the eye is predominantly one of IgG production, and since these cells within the eye during the primary reaction presumably provide the seed of immunologic memory for the booster response to systemically administered antigen,1 it is not surprising that the booster response within the eye is also predominantly characterized by IgG production. The predominant distribution of cells with a chain within the conjunctiva correlates very well with the earlier observation15 of the presence of appreciable amounts of IgA within this tissue. Almost all of the a chain-containing cells found were distributed within the conjunctival stroma, although occasional cells were found infiltrating the epithelial layers. The close agreement of our values for the proportions of the various heavy chain-containing cells accords well with values reported earlier for lymphoid tissues7-s and for other secretory tissues,1-'- 1C~1S and appears to confirm the general validity of the methods employed in the present experiments. A single intravitreal injection of OA did not appear to have any effect on the cells of the preauricular lymph nodes. There was no significant difference in the proportion of heavy chains between the ipsi- and contralateral preauricular nodes, nor were specific antibody-containing cells found in either of these lymph nodes or in the perilimbal conjunctiva. In fact, only


a few specific anti-OA-containing cells were found within the spleen early during the primary response. These findings indicate that the antigen injected into the vitreous does not find its way in sufficient amounts to stimulate antibody formation in the conjunctiva or preauricular nodes, and probably enters the circulation to induce only a minimal stimulus in the spleen itself.19 Perhaps the most fascinating aspect of the data reported in Tables I to IV relates to the specificities of the antibodies formed within the eye after the induction of specific immunogenic uveitis. It is almost an article of immunologic faith that the driving force for differentiation of a lymphoid cell to antibody formation is the specific antigen itself,20-21 acting either alone or in concert with thymus-derived cells22 or their products.23"25 It is, therefore, not surprising to find only a small proportion of Ig-cells within the spleen and lymph nodes forming antibody specific for the antigen employed in these experiments, since these tissues are normally exposed to a variety of extraneous antigenic stimuli. However, the anatomic isolation of the ocular tissues suggests that they should be fairly well restricted to a specific response to the antigen introduced within the eye. If we consider together the tissues of the uveal tract and retina, then we find that early during the course of the primary reaction to intravitreal antigen, only one of 545 Ig-forming cells appears to be making specific anti-OA antibody (Table I). Similarly, late in the primary reaction only seven per cent of Ig-cells are found to make specific antibody, while early in the secondary booster response some four per cent of cells are specifically involved, and late in the secondary booster response fewer than 18 per cent are responding to the homologous antigen. Thus, under the most favorable of circumstances, some 80 to 90 per cent or more of the antibodyforming cells within the eye appear to direct their activities against antigens other than that used to incite the immunologic


response (making the now-reasonable assumption that all immunoglobulin is antibody against something). But this observation of antibody formation within the eye is by no means unique. It is a common experience that hyperimmunization of animals generally results in a much larger increase in total immunoglobulins than can be accounted for in terms of antibody specific for the antigen employed. Again, in a simpler system, we have observed20 that immunization of the immunologically virgin and agamma-globulinemic fetal lamb in utero with OA yields a response in which specific anti-OA antibody makes up something less than one per cent of the total immunoglobulin formed in response to this stimulus. It is well known that certain B-cell mitogens, such as lipopolysaccharide, may nonspecifically stimulate plasma cell precursors to the formation of a polyclonal antibody response.27 It has recently been suggested that the specific interaction of an antigen with sensitized lymphocytes might lead to the release of effector substances which might act nonspecifically in the same manner.2*5' 29 This latter possibility is currently under investigation, since such nonspecific mechanisms might help to clarify some of our current conceptual problems in defining the etiology and pathogenesis of recurrent anterior uveitis.20' 30 REFERENCES 1. Zimmerman, L. E., and Silverstein, A. M.: Experimental ocular hypersensitivity: histopathologic changes observed in rabbits receiving a single injection of antigen into the vitreous, Am. J. Ophthalmol. 48: 447, 1959. 2. Smith, R. E., Jensen, A. D., and Silverstein, A. M.: Antibody formation by single cells during experimental immunogenic uveitis, INVEST. OPHTHALMOL. 8: 373,

1969.

3. Hall, J. Ml, and O'Connor, G. R.: Correlation between ocular inflammation and antibody production. II. Hemolytic plaque formation by cells of the uveal tract, J. Immunol. 104: 440, 1970. 4. Silverstein, A. M.: Ectopic antibody formation in the eye: pathologic implications, in: Immunopathology of Uveitis, Maumenee, A.


Volume 14 Number 8

E., and Silverstein, A. M., editors. Baltimore, 1964, Williams & Wilkins, p. 83. 5. Eifrig, D. E., and Prendergast, R. A.: Anterior chamber lymph node implantation: Local adoptive immune response in the eye, INVEST. OPHTHALMOL. 7: 293,

1968.

6. Cohen, S., and Porter, R. R.: Structure and biologic activity of immunoglobulins, Adv. Immunol. 4: 287, 1964. 7. Bernier, C , and Cebra, J. J.: Frequency distribution of a, y, K and X polypeptide chains in human lymphoid tissues, J. Immunol. 95: 246, 1965. 8. Cebra, J. J., Colberg, J. E., and Dray, S.: Rabbit lymphoid cells differentiated with respect to «-, y-, and /i-heavy polypeptide chains and to allotypic markers AI and A2, J. Exp. Med. 123: 547, 1966. 9. Cebra, J. J., and Goldstein, C : Chromatographic purification of tetramethyl-rhodamineimmune globulin conjugates and their use in the cellular localization of rabbit y-globulin polypeptide chains, J. Immunol. 95: 230, 1965. 10. Jaquet, H., Bloom, B., and Cebra, J. J.: The reductive dissociation of rabbit immune globulin in sodium dodecylsulfate, J. Immunol. 92: 991, 1964. 11. Wood, B. T., Thompson, S. H., and Goldstein, G.: Fluorescent antibody staining. III. Preparation of fluorescein-isothiocyanate labeled antibodies, J. Immunol. 95: 225, 1965. 12. Crandall, R. B., Cebra, J. J., and Crandall, C. A.: The relative proportions of IgG-, IgA-, and IgM-containing cells in rabbit tissues during experimental trichinosis, Immunology 12: 147, 1967. 13. Silverstein, A. M., Welter, S., and Zimmerman, L. E.: A progressive immunization reaction in the actively sensitized rabbit eye, J. Immunol. 86: 312, 1961. 14. Silverstein, A. M.: Effect of x-irradiation on the development of immunogenic uveitis, INVEST. OPHTHALMOL. 2: 58,

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

1963.

15. Allansmith, M., and Hutchison, D.: Immunoglobulins in the conjunctiva, Immunology 12: 225, 1967. 16. Tomasi, T. B., and Bienenstock, J.: Secretoiy immunoglobulins, Adv. Immunol. 9: 1, 1968. 17. Brantzaeg, P., Fjellanger, I., and Gjeruldsen, S. T.: Localization of immunoglobulins in human nasal mucosa, Immunochemistry 4: 57, 1967. 18. Crabbe, P. A., Carbonara, A. O., and Heremans, J. F.: The normal human intestinal mucosa as a major source of plasma cells

29.

30.

containing y-A immunoglobulin, Lab. Invest. 14: 235, 1965. Fernando, A. N.: Immunological studies with I i a i labeled antigen in experimental uveitis, Arch. Ophthalmol. 63: 515, 1960. Wilson, J. D., and Feldmann, M.: Dynamic aspects of antigen binding to B cells in immune induction, Nature 237: 3, 1972. Rajewsky, K., Schirrmacher, V., Nase, S., et al.: The requirement of more than one antigenic determinant for immunogenicity, J. Exp. Med. 129: 1131, 1969. Miller, J. F. A. P., Sprent, J., Basten, A., et al.: Cell-to-cell interaction in the immune response. VII. Requirement for differentiation of thymus-derived cells, J. Exp. Med. 134: 1266, 1971. Sjoberg, O., Andersson, J., and Moller, G.: Reconstitution of the antibody response in vitro of T-cell-deprived spleen cells by supernatants from spleen cell cultures, J. Immunol. 109: 1379, 1972. Feldmann, M., and Basten, A.: Cell interactions in the immune response in vitro. IV. Comparison, of the effects of antigen-specific and allogeneic thymus-derived cell factors, J. Exp. Med. 136: 722, 1972. Rosenthal, M., Stastny, P., and Ziff, M.: Stimulation of y-globulin synthesis and specific antibody production by a factor released by activated lymphocytes, J. Immunol. I l l : 1119, 1973. Silverstein, A. M., Thorbecke, G. J., Kraner, K. L., et al.: Fetal response to antigenic stimulus. III. y-Globulin production in normal and stimulated fetal lambs, J. Immunol. 91: 384, 1963. Andersson, J., Moller, G., and Sjoberg, O.: Selective induction of DNA synthesis in T and B lymphocytes, Cell. Immunol. 4: 381, 1972. Rubin, A. S.. and Coons, A. H.: Specific heterologous enhancement of immune responses. II. Immunological memory cells of thymic origin, J. Exp. Med. 135: 437, 1972. Hall, J. M., and Pribnow, J. F.: Specificity of antibody formation after intravitreal immunization with bovine gamma-globulin and ovalbumin. II. Nonspecific enhancement of the secondary response, Cell. Immunol. 11: 64, 1974. Silverstein, A. M.: Immunogenic uveitis: Wherein it is demonstrated that the cure may sometimes be worse than the disease, Trans. Ophthalmol. Soc. U. K. In press.