Journal o f Clinical Immunology, VoL 10, No. 6 (November Supplement 1990)

The Use of Monoclonal Antibodies for Treatment of Autoimmune Disease L A W R E N C E STEINMAN 1

monoclonal antibody reagents. Other approaches to treatment of autoimmune disease based on targeting the variable region of the T-ceU receptor involve active molecular vaccination.

Over the past decade monoclonal antibodies have been successfully employed in a number of animal models of autoimmune disease. We have used antibodies to the class II gene products of the major histocompatibility complex, the CD4 molecule on helper T cells, and the T-cell receptor. Monoctonal anti-class II antibodies have been administered to treat paralytic disease in the animal model of multiple sclerosis---experimental allergic encephalomyelitis. These antibodies not only reverse acute paralytic disease but also decrease the number of relapses in a model of relapsing/remitting multiple sclerosis when given after the first attack. The advantage of this form of therapy is that it is haplotype specific. In other words, in a heterozygous individual it is possible to block the major histocompatibility gene associated with disease susceptibility while leaving other major histocompatibility gene products free for antigen presentation. Thus, animals given this form of immunotherapy are not significantly immunosuppressed. Antibodies to the CD4 molecule have been equally effective in treating animal models of autoimmunity. We and others have reversed ongoing paralysis in experimental autoimmune encephalomyelitis. Relapses have been diminished after the administration of anti-CD4. Antibodies to CD4 have been used successfully to treat animal models of systemic lupus erythematosus, rheumatoid arthritis and myasthenia gravis. Recent trials with anti-CD4 have been successful in the treatment of rheumatoid arthritis and cutaneous T-ceU lymphoma. The latter trial employed a chimeric human/mouse antibody. Antibodies to the variable region of the T-cell receptor have been employed to treat experimental autoimmune encephalomyelitis. These antibodies were effective in both preventing and reversing ongoing disease. These antibodies targeted the variable region gene products of T-cell receptors that were involved in autoimmune disease. It is remarkable that a limited heterogeneity of T-cell receptors is responsible for autoimmune conditions. However, in certain instances the T-ceU receptor repertoire is more diverse and may require a cocktail of

KEY WORDS: Experimental allergic encephalomyelitis (EAE); major histocompatibility complex (MHC); multiple sclerosis (MS); myelin basic protein (MBP); T-cell receptor (TCR).

The focus of this article is the T-cell receptor (TCR), that part of the immune system which acts in conjunction with the major histocompatibility complex (MHC), in recognizing an autoantigen. The TCR does not recognize native protein but recognizes a fragment of the antigen, often a contiguous portion of sequential amino acids, which is bound to the MHC. Recognition occurs through both class I MHC antigens and class II MHC antigens. The emphasis here is on the class II MHC recognition because it is quite significantly involved in a number of autoimmune diseases. Many diseases that are either proven or suggested to be autoimmune are associated with human leukocyte antigen (HLA) class II molecules, primarily HLA-DR (Table I). The first part of this paper is a review of monoclonal antibody therapy. Later an exciting approach that may be competitive with monoclonal antibody therapy, namely, the use of peptide molecules to block T-cell recognition by the MHC, is discussed. The relative advantages of the two approaches are compared and contrasted. Before discussing human studies, I want to familiarize the reader with an animal model that has been used in my laboratory for the last decade to delineate how monoclonal antibodies and peptides can be used immunotherapeuticalty. This model, experimental allergic encephalomyelitis (EAE), is very important in the history of modem immunology. Patients who were vaccinated with the original

1Department of Neurologyand Genetics, Stanford University, Stanford, California94305. 30S

0271-9142/90/1100-030S$06.00/0© 1990PlenumPublishingCorporation

MONOCLONAL ANTIBODIES FOR AUTOIMMUNE DISEASE

31S

Table I. H L A A s s o c i a t i o n with M y a s t h e n i a Gravis and Other A u t o i m m u n e D i s e a s e s Is Localized to Discrete Regions of Allelic Variability"

Disease

Susceptibility

H L A association

Localized a m i n o acid s e q u e n c e in the ~31 d o m a i n 65

Increased

KDLLEQKRAA ...... R- - ...... R- - ...... R--. . . . . RR- - -

Normal/decreased

DR4-Dwl0

--I--DE---

51 Diabetes mellitus (IDDM)

56

60

Increased

DR4/DQw3.2 DR3/DQw2 DR1/DOwl.I DR2/DOwI.AZH DR4/DOw3.1 DR5,~Qw3.1

TPLGPPAAEY -L- -L ..... --Q-R-V----Q-R-S--...... D--...... D---

Normal/decreased

DR2/DQwI.12 DR2/DQwl.2

--Q-R-D----Q-R-D--51

Absolute

DRw6/DQB1.9

Increased

DR3/DQw2

56

60

TPQGRPDAEY

51 M y a s t h e n i a gravis (MG)

74

DR4-Dw4 -Dwl4 -Dwl5 DR1-Dwl DRwl0

R h e u m a t o i d arthritis (RA)

P e m p h i g u s vulgaris (PV)

70

56

60

TLLGLPAAEY

Data s u m m a r i z e d from T o d d et aL (2).

rabies vaccine of Louis Pasteur--and did not become infected with rabies--often developed paralytic accidents due to contamination of the rabies vaccine which had been grown in rabbit spinal cord cultures. In a series of experiments, aimed at understanding these complications of the Pasteur treatment, done primarily at the Rockefeller Institute by Rivers and Freund and at Columbia University by Kabat, E A E came to be understood as an autoimmune attack against components of the central nervous system, specifically myelin and some of the major myelin proteins. Experimental allergic encephalomyelitis served as a model for complications of the Pasteur vaccine for rabies. This disease could be induced in mice, rats, and monkeys, and Freund's adjuvant contributed to the high frequency of the disease. The EAE model, developed by some of the pioneers in immunology, has turned out to be very useful for working out strategies for modern immunotherapy. Although EAE does not represent multiple sclerosis (MS), it does show

some parallels to the human demyelinating diseases. One advantage of this model is the obvious readout. When a mouse is paralyzed with EAE its legs drag, its tail droops, and its head hangs down. Clearly, any therapy that can take an animal from this state to a healthy state is of great interest. In recent years it has been recognized that antigen-specific cloned T cells can be used to induce EAE. If one introduces a large number of T cells that are capable of recognizing myelin, the animal will develop a rapidly fatal disease reminiscent of the worst cases of acute disseminated encephalomyelitis in humans. With short latency the animal goes on to develop fulminant, grade 5, paralysis (on a scale of i to 5, 5 is moribund) and dies. However, with a smaller inoculum of these same cloned T cells, one can induce a disease with a longer latency that relapses and remits (Fig. 1). This is fascinating because many of the human autoimmune diseases often have their onset in young adulthood and have the capacity to relapse and remit. The regulatory

Journal of Clinical Immunology, Vol. 10, No. 6 (November Supplement 1990)

32S

STEINMAN

5

-

3~ 5 1

5

t

1

Vi

5 c-

z ....

I

C

i

D

1 5

E

1

35 E-~rt 1

I F i

l

i

I

I

~

i

i

t

i

30 60 90 Days after Injection of PJR-25 Fig. 1. Clinical observation of individual (PLSJ)F 1recipient mice following injection of T-cell clone PJR-25. Representative mice from each of the following groups are shown: A 005 cells); B and C (5 x 105); D and E (106); F (5 x 106). Clinical disease was graded as follows: 0, no observable signs of clinical disease; I, less of tail tone only; 2, mild paraparesis; 3, moderately severe paraparesis; 4, complete paraplegia; 5, moribund; ¢, death.

mechanisms that allow for this relapsing/remitting condition are an exciting area for study. It should be stressed that it is simply the difference in the size of the inoculum that can turn an acute fatal disease in mice into a relapsing/remitting disease with a relatively long latency. While some imagination must be used to convert from a scale of days in animals to years in humans, the analogy is tantalizing. In the animal model we see infiltration of lymphocytes in the perivascular spaces of the brain. The lymphocytes are predominantly CD4 positive, although CD8-positive lymphocytes and some B cells can be found. All the players in the cellular immune system are present. Cloned T cells can cause fulminant demyelination in our model and control clones are unable to cause that type of histological problem. Another important point is the aberrant expression of the MHC molecules. We do not normally think of the brain as an organ that has much expression of MHC class I and certainly it has no expression of MHC class II. However, if one uses an immunoperoxidase antibody system, a considerable amount of MHC class II can be seen around the perivascular cuff. Counterstaining these cells with glial fibrillary acidic protein shows that they are not normal components of the immune system; they are astrocytes. This underscores the point that the expression of MHC class II is highly inducible and susceptible to regulatory molecules such as

gamma interferon. This inducible expression on cells, such as astrocytes that are not normally part of the immune system, may be absolutely essential in other diseases such as rheumatoid arthritis, where cells in the synovium can become potent M H C class II expressers (1, 2). The potential site of action for some of these immunotherapies may have to be not only peripheral immune organs such as lymph nodes, spleen, and thymus but also the target organ, in which the antibody may have to interact with cells from other organ systems. A number of neurotransmitter substances and hormones can influence this inducible regulation of MHC class II, hence there can be interaction of different organ systems with the immune system, e.g., endocrine systems and neurological systems. In 1981, in collaboration with Hugh McDevitt, we showed that monoclonal antibodies to M H C class II were capable of preventing E A E (3). We have also used the relapsing/remitting model and began treatment of mice after they had suffered the first attack of paralysis (4). Over a few months we were able to show that we could significantly reduce the number of further relapses and reduce mortality. When we treated with the appropriate anti-MHC class II antibodies, the number of relapses was reduced and mortality was dramatically reduced. So in this model disease monoclonal antibodies could be useful for both prevention and therapy. We have also shown that this can work in nonhuman primates. These antibodies have not shown toxicity. In collaborative work done in the Netherlands we showed that anti-DQ antibodies were capable of treating EAE in rhesus monkeys, as were anti-DR antibodies. With no treatment these monkeys die within 2 days after developing paralysis. We had very longterm survivors; the experiment was terminated at 200 days (5). Before considering antibodies to TCRs more information is given about this animal model to show what can come from such studies. After eliciting EAE in highly inbred strains of mice, we can map the exact specificity of the T-cell response to myelin so as to investigate what genes of the TCR variable region are involved in that recognition. We have used two strains of mice, the PL/J and the S J L The PL/J mouse recognizes a fragment of myelin basic protein (MBP) at the N terminus between amino acid 1 and amino acid 11, while the SJL mouse recognizes a fragment between residue 89 and residue 101. Interestingly the F1 animal mimics the PL/J parent b y responding mainly to the N-terminal

Journal of Clinical Immunology, Vol. 10, No. 6 (November Supplement 1990)

MONOCLONAL ANTIBODIES FOR AUTOIMMUNE DISEASE

Table II. Discrete T-Ceil Epitopes of MBP in Mice Peptide

Encephalito- Class II genicity restriction

TCR ~ V~ usage

Ref. No.

AcN1-20

+

I-A~A~ V~8 predominantly

6

N1-20

-

I-A,~A~ V138predominantly

7

AcN9-16

-

I-A~A~ V~8 predominantly

6

I-A~A~ Not known

6

I-E~,E~ Not V~8

7

1-E~E~

7

N35-47

+

Not V~8

N89-101

+

I-A~A~ V~17 predominantly

8

N89-100

+

I-A~A~ Not V~17

8

aT-cell antigen receptor.

fragment instead of having codominant expression of this phenotype, demonstrating susceptibility to disease with either fragment of myelin. This has implications for some genetic studies suggesting that even if one understands the genotypes very well, the phenotype often is modified by factors that are not predictable by mendelian genetic analysis. We have mapped the various portions of the myelin molecule which we know can cause paralytic disease in these mice (Table II) (6-8). Fragment 1-11 is important in the PL/J mouse and is restricted by the class II molecule I-A ~ in the alpha chain and I-A u in the beta chain. I-A is the homologue of MHC class II molecules in the human. The I-E molecule is similar to HLA-DR and the I-A molecule is similar to HLA-DQ. We have mapped other myelin fragments: 35-47 is restricted by the I-E molecule and 89-101 is the fragment that causes EAE in SJL mice. We took a number of clones that are capable of causing EAE and sequenced their TCRs. These are clones that recognize the 1-11 amino acids of MBP. We found that seven of eight T-cell clones that were capable of causing EAE used the same TCR of the V beta gene, and seven of eight used the same TCR of the V alpha gene (9). The conclusions of Hood and his

33S

group paralleled ours; there was predominant usage of the V beta 8 molecule and predominant usage of a V alpha molecule (10). However, Hood's group used a different strain of mouse (B10.PL), which has the same M H C as the PL/J mouse but a slightly different TCR repertoire. This congenic mouse used the same V beta but a different V alpha in recognizing the 1-11 fragment of MBP. The sequences that H o o d ' s group and ours mapped at the junctional region between the variable and the joining regions of the TCR show a considerable degree of conservation. There is conservation in the nucleotide sequences and in the amino acid-derived sequences at the junctions, for both the V alpha and the V beta regions. In 1988 Hood's group and we showed that animals paralyzed b y EAE could be completely cured by administering a monoclonal antibody directed against the TCR V beta 8, a gene product that is involved in recognition of the peptide associated with EAE (9, 10). If one compares a treated animal with an untreated control, the difference is remarkable. One can begin to see amelioration of the disease within even 1 day. It is apparent even to the blinded observer that something important is happening. We also induced the disease with the whole protein which contained multiple fragments capable of causing EAE, in other words, multiple epitopes of protein fragments 1-11, 35-47, and 89-101, but there was a dominant reaction to 1-11 in this animal model. When we eliminated all or most of the T cells that were capable of reacting to fragment 1-11, we were able to reduce the number of paralyzed animals within 72 hr (Table III). These results continued at 14 days and even at 2 months. To summarize what has been discussed so far, in the PL/J and the SJL mouse there is a very limited heterogeneity of the TCR V elements. For each of the V beta, J beta combinations, there was a specific amino acid sequence in the junctional region.

Table III. Reversal of Guinea Pig MBP-Induced Disease with V~8-Specific Monoclonal Antibody F23.1 a Number of mice with clinical symptoms 14 days after treatment

Number of mice with clinical symptoms 72 hr after treatment Treatment F23.1 KJ23a

None

Mild

Severe

None

Mild

Severe

Deaths

12 1

5 12

2 9

14 9

3 2

1 7

1 4

~Disease was induced with guinea pig MBP. Treatment was started 24 hr after the first clinical signs of EAE were detected. The experiment was done in a double-blind fashion. The mice received 200 ~xg of the antibodies i.p. Monoclonat antibody F23.1 depletes the V~8-positive T cells and KJ23a depletes the V~tV-positive T cells in vivo. Both antibodies deplete similar amounts of T cells in the (PLxSJL) F~ mice used in these experiments. Data summarized from Acha-Orbea et at.

Journal of Clinical Immunology, VoL 10, No. 6 (November Supplement 1990)

34S

STEINMAN

Differences in the junctional regions for the beta chain and the alpha chain did not lead to differences in fine specificity. We can both prevent and treat this model autoimmune disease (EAE) with antibodies to the TCRs. Recent articles in N a t u r e (11) and in Science (12) show that if the peptides are removed from those junctional regions and mice are vaccinated, EAE can be prevented. If we knew what TCRs were involved in autoimmune diseases, we might be able to intervene in the acute stage of diseases with specific antibodies and later treat individuals with a molecular vaccine against some of their TCR junctional combinations. This concept assumes the availability of humanized or chimeric antibodies to overcome problems with antibody rejection. It is a very exciting prospect that seems feasible on the basis of animal models. On a less optimistic note, it may sometimes be necessary to use antibody cocktails. Indeed, the autoimmune response to certain peptide fragments of myelin is not absolutely restricted. In some cases, the immune system behaves in a very devilish fashion when an autoimmune disease develops. We saw a dramatic example of this when we moved from PL/J mice to SJL mice. We knew the fragment that the mouse responded to was 89-101; T-cell clones that use that particular TCR gene, called V beta 17, could respond to 89-101. However, as soon as we removed the proline from position 101 in MBP, those clones would not recognize it, and yet within the same animal there were other clones which used TCR genes other than V beta 17 that could respond to either 89-101 or 89-100. When we induced disease with clones that were restricted to the use of the V beta 17 gene, in other words, those clones that would recognize only 89-101, the mice were paralyzed. If we used an antibody to the product of that V beta 17 gene, we could block disease. However, if we induced disease with clones that were V beta I7 negative, i.e., clones that could recognize either 89-101 or 89-100, our antibody treatment was ineffective. That tells us that a cocktail of antibodies may be required to eliminate TCRs other than V beta 17. This is an example of an extremely fine-tuned response to the truncation of one amino acid. In antigen processing or just in the normal proteolysis of exposed cell protein fragments, one could imagine the intensification of an autoimmune disease against nested epitopes. The attack on these nested portions of autoantigens could lead to a vast distribution of TCRs involved in

autoimmunity and make therapy with a monoclonal antibody an unlikely proposition (8). The next point for discussion is the use of peptides as immunotherapeutics to block the MHC. What MHC molecules are involved? What selfantigen fragments, what peptides of autoantigens might be involved? Our group, in collaboration with Hugh McDevitt, published the first demonstration of the success of this approach (13, 14). In the same issue of Cell there is an article b y Hood and colleagues which shows essentially the same results (15). A whole industry has sprung up. Start-up companies such as Immulogics, in Cambridge, Massachusetts, and Cytel, in La Jolta, California, and pharmaceutical companies including Sandoz and Ciba-Geigy are very excited about this approach; but even though we have been able to make it work, I believe that it has major limitations. Using X-ray crystallography Wiley et al. and Bjorkman have shown that the M H C molecule contains a binding cleft which houses a fragment of the protein antigen. (I am extrapolating somewhat in going from what was actually shown to what may be happening with class II molecules, but it is probably correct.) This binding cleft measures about 10x 10x20 A and it can house a peptide consisting of 10 to 20 amino acids if the peptide is packed in an alpha helical or amphipathic helical configuration. This cleft can house the kind of peptide described previously in the EAE study. Around the floor and walls of this cleft are the highly polymorphic residues of the MHC molecule, including some of the amino acids that have a high degree of variability on Kabat Wu plots, which undoubtedly play an important role in susceptibility to certain autoimmune diseases. The concept is to make a peptide that could displace or somehow block the fragment of self-antigen from getting into that MHC cleft and triggering a T celt. The ultimate hope is to create a compound with the same structural features as the peptide that wilt not be chewed up by peptidases. It may be possible. In the EAE model we knew everything about the peptide specificity and the I-A restriction so we could devise some peptide-blocking agents and show in vitro that a good competitor could be made. For instance, we found out that acetylation is needed at the N terminus and we know that MBP is normally acetylated. What happens if a peptide without the acetyl group, such as fragment 1-20, is introduced? If the EAE is induced with the 1-11 fragment, approximately half of the animals become paralyzed. If nonacetylated 1-20 is used, the dis-

Journal of Clinical Immunology, Vol. ]0, No. 6 (November Supplement 1990)

MONOCLONALANTIBODIESFOR AUTOIMMUNEDISEASE ease is prevented. We found other blockers in addition to nonacetylated 1-20. The part of the molecule between amino acid 9 and amino acid 20 was not capable of causing paralysis and it could compete effectively with 1-11 (14). Something even more mysterious was observed in one of our studies. We were conducting mutagenesis of the peptide by inserting alanine at each and every position to see what effect that would have on the ability of 1-11 to trigger EAE. When an alanine was substituted for a lysine at position 4, we observed what is called a heteroclitic response, meaning that the unnatural molecule responded thousands of times better than the natural molecule. We thought that if that molecule caused a response that was so much better than the T-cell done, it would probably cause super disease. We expected the animals to become sick and be paralyzed sooner than usual. To our surprise, this mutagenized molecule turned out to induce tolerance and did so very efficiently. Questioning how this mutagenized molecule was working, we used a technique developed by Emil Unanue and Paul Allen and demonstrated direct binding of the mutagenized molecule to MHC. Normally if one introduces a labeled molecule and tries to compete with the natural molecule, the natural antigen cannot be driven out until there is an excess of about 1000-fold. That holds true when alanine is inserted at position 3; the natural molecule can be driven out at 1000x but not at 100x. In other words, the mutation we had made in the peptide caused it not only to bind exceptionally well to the MHC but also to induce tolerance. We were able to prevent EAE completely in two experiments using this modified peptide (13). How it works and whether the phenomenon applies generally will be interesting to see. It may be that this "superbinder" is overstimulating the T cell, creating tolerance. Based on studies of this type we can now say that residues 1, 3, and 6 are in contact with the TCR, while residues 2, 4, and 5 are probably in direct contact with the MHC. We can begin to think on a very detailed level about how such molecules might be modified to make them bind not to the MHC but to the TCR. Alternatively, we may want to modify them so that they still bind to the MHC but are unable to interact with the TCR. This suggests a whole level of therapeutics that may some day be available to prevent autoimmune disease. Interestingly, neither we nor Hood's group can reverse autoimmune disease with this approach, in marked

35S contrast to what we see with monoclonal antibodies. Blocking T-cell recognition may be good for prevention or to block further disease induction and further deterioration in individual patients, but I believe that monoclonal antibodies will have greater use for acute therapy in reversing ongoing disease. When a patient is developing blindness or paralysis, one does not want to wait around and hope to vaccinate him with the junctional peptide or feed in a blocking agent to block his MHC. One wants to get rid of those T cells as soon as possible and monoclonal antibodies will be very useful for performing that function. Table IV summarizes the various kinds of therapy that we may be looking at in the future. These include vaccination with peptides of the TCR and infusing antibodies against TCR, MHC class II, or CD4. It was mentioned previously that the MHC is associated with a number of autoimmune diseases. With the advent of the polymerase chain reaction, we now have a vast amount of knowledge, derived primarily from Hugh McDevitt's laboratory at Stanford and Henry Erlich's laboratory at Cetus. These investigators have studied a number of patients with autoimmune diseases to determine if there is anything different about their MHC molecules that would distinguish them from others who do not have autoimmune disease. Certain very striking examples of changes in the MHC gene are associated with autoimmune disease. The two that are most striking relate to the change at codon 57 in the HLA-DQ beta chain. One example is in a very common human disease, insulin-dependent diabetes mellitus (IDDM). Certain codon changes at the DQ beta 57 position are absolutely critical in conferring either susceptibility or resistance to IDDM (16). I was fortunate to be involved in a study of another disease, an autoimmune disease of skin called pemphigus vulgaris, in which individuals make antibodies against a cement substance that holds the skin together. We found a codon change, again at position 57, that conferred almost absolute susceptibility to the disease if the individual carried a certain mutation (17, 18). This is the clearest example of a mutation in a codon that confers almost absolute disease susceptibility. If one has the mutation, one is going to develop pemphigus vulgaris. Patients who make antibodies against the substance that holds skin together will suffer from chronic blistering and infection. The codon change at residue 57 is extremely unusual in normal populations. This study was carried out in two places, in

Journal of Clinical Immunology, VoL I0, No. 6 (November Supplement 1990)

36S

STEINMAN

Table IV. Experimental Autoimmune Disease Can Be Treated by Disruption of T-Celt Recognition Disease prevented by treatment with T-cell vaccination Experimental autoimmune encephalomyelitis (H-2u mouse) Experimental autoimmune encephalomyelitis (H-2s mouse) Experimental autoimmune encephalomyelitis (Lewis rat) Experimental autoimmune myasthenia gravis (mouse) Experimental autoimmuune thyroiditis (mouse) Nonobese diabetic (mouse) Collageninduced arthritis (mouse)

Anti-CD4

Blocking peptides

Anti-TCR

Anti-Class II

NT~

Yes (3)b

Yes (6)

NT

Yes (18)

NT

No (4)

Yes (7)

Yes (12)

NT

Yes (1)

Yes (5)

NT

Yes (13)

NT

NT

NT

Yes (8)

Yes (14)

NT

Yes (2)

NT

Yes (9)

NT

NT

NT

NT

Yes (10)

Yes (15, 16)

NT

NT

NT

Yes (11)

Yes (17)

NT

"Not tested. bReferences as follows: (1) Lider et aL: Science 239:181-183, 1988. (2) Maron et aL: J Immunol 131:2316-2322, 1983. (3) Acha-Orbea et aL: Ceil 54:263-273, 1988. (4) Sakai et aL: Proc Natl Acad Sci USA 85:8608-8612, 1988. (5) Owhashi and Heber-Katz: J Exp Med 168:2153-2164, 1988. (6) Steinman et aL: Proc Natl Acad Sci USA 78:71tl-7114, 1981. (7) Sriram et aL: J Immunol 139:1485-1489, 1987. (8) Waldor et aL: Proc Natl Acad Sci USA 80:2713-2717, 1983. (9) Vladutiu and Steinman: Cell Immunol 109:t69-180, 1987. (10) Boitard et aL: Proc Natl Acad Sci USA 85:9719-9723, 1988. (11) Wooley et al.: J Immunol 134:2366-2374, 1985. (12) Waldor et aL: Science 227:415-417, 1985. (13) Brostoff and Mason: J Immunol 133:1938-1942, 1984. (14) Christadoss and Dauphinee: J Immunol 136:2437-2440, 1986. (15) Wang et al.: Diabetes 36:535-538, 1987. (16) Koike et aL: Diabetes 36:539-541, 1987. (17) Ranges et aL: J Exp Med 162:1105-1110, 1985. (18) Sakai et aL: Proc Natl Acad Sci USA 86:9470-9474, 1989.

Israel among both Ashkenazi and Sephardic Jews and in Vienna, Austria, among non-Jews. The same constellation held up, namely, that a mutation of codon 57 was important in conferring susceptibility to the disease. Codon 57 lies at a very important portion of the MHC-binding cleft which determines a contact site with the putative antigen in this disease. It is also in a fortuitous position for TCR interaction with that antigen. Thus, it is probably no mystery that this particular mutation is involved with the development of a dreadful disease. We may be able to target M H C molecules with blocking peptides or with antibodies but we will run the risk of eliminating particular M H C mole-

cules that are very important in normal immune function. Except for pemphigus vulgaris, diseaseassociated M H C molecules are found in both the normal and the sick, and even those individuals who are sick are probably using M H C molecules for important protective immune responses. It could be argued that if one is heterozygous for HLA-DR, DQ, and DP, eight different combinations of MHC molecules are possible. Furthermore, there may be additional variability because a DR alpha chain from one parent could be combined with a DR beta chain from another parent, increasing the possible combinations to 16 or maybe 32. If only the one associated with disease

Journal o f Clinical Immunology, Vol. 10, No. 6 (November Supplement 1990)

MONOCLONAL ANTIBODIES FOR AUTOIMMUNE DISEASE

is eliminated, the rest of the M H C molecules will be intact and therefore the individual would not be significantly immunosuppressed. We do not have enough information to know whether that argument is valid. As for TCRs and diseases, there are about 20 to 25 V alpha families and 25 V beta families in humans. A number of investigators are looking at a number of diseases to evaluate whether there is an overexpression of the V alphas or V betas that are associated with that disease in the germ line and perhaps at the site of the autoimmune attack. Ultimately we would hope to reach the point we have reached in EAE and be able to tell which TCRs are involved in the autoimmune attack and make reagents to destroy or block that particular V alpha or V beta. There is the hope that taking away one V beta or one V alpha would not adversely affect the important immunity to infectious pathogens. It is only a hope at present but in my view it is realistic. Our present study is focused on MS because, as a clinical neurologist, I see many patients with this disease. We have observed that a particular V alpha family, V alpha 12, occurs more commonly in the germ line of MS patients than in others; however, V alpha 12 is seen also in healthy individuals. Using a Southern blot analysis we can demonstrate that the V alpha family is transmitted as a mendelian dominant trait through families (19). We studied patients with MS in northern California, Australia, and Japan. We found that there is a certain polymorphism in this V alpha family that is associated with MS in about 70 to 80% of patients in northern California. Another band shows almost 90% association in northern California, Australia, and Japan. These data suggest that this polymorphism in the V alpha somehow is associated with MS. We have gone further and shown that the polymorphism is not in the coding region; it is in some flanking region. We asked the question, Is this V alpha gene rearranged in the MS lesion and expressed in the lesion? The answer is yes. We can take the small amount of messenger RNA obtainable at autopsy from a region of the brain where there is a demyelinating plaque and, using the polymerase chain reaction, show a productive rearrangement in the TCR. We know that the V alpha 12 rearranges in the brain (20). There are some human autoimmune diseases in which we know the complete structure of the antigen involved. One of them is myasthenia gravis, a disease in which the patient makes antibodies to

37S

acetylcholine receptor (AchR). We have attempted to determine which fragments of the AchR myasthenics mount responses to. Antibodies acting on the AchR lead to the blockade of neuromuscular transmission and an anatomical change in that exquisite postsynaptic folding of the muscle membrane that allows for neurotransmission. Numa has sequenced the AchR and we have made peptides from various regions. We know that individuals who are HLA-DR3 positive are making T-cell responses primarily against peptide 257-269 and myasthenic individuals who are HLA-DR5 positive are making responses to 195-212 (21). Knowing the antigen and something of the specificity of the response, one can begin to devise the type of mutant peptides that were successful in our studies on EAE. Myasthenia gravis is just one example of what could be done in autoimmune disease by looking at one portion of the immune response, i.e., the one involving T cells restricted by the MHC class II. In devising a therapy we have to be knowledgeable about the disease, its pathogenesis, and the clinical situation. If the pathogenesis of the disease allows for it, the ultimate therapy might be molecular vaccination against TCRs. If there really is restricted usage, then it is conceivable that 50 years from now the routine childhood immunizations would include immunizations not only against infectious pathogens but also against portions of the individual's own immune system. That is science fiction now but it may come into being. In the meantime, we certainly do want to pay attention to other modalities of therapy, anti-Tac, anti-CD4, and anti-MHC, and use them in a rational way. We just do not know enough about the pathogenesis of autoimmune diseases to know whether the extremely elegant and specific approaches which were described will actually work or whether the immune system will be so vicious and devilish that we will be limited to using more nonspecific therapy. We should develop both the highly specific and the less specific therapies.

REFERENCES 1. Wraith DC, McDevitt HO, Steinman L, Acha-Orbea H: T cell recognition as the target for immune intervention in autoimmune disease. Cell 57:709-715, 1989 2. Todd JA, Acha-Orbea H, Bell JI, Chao N, Fronek Z, Jacob CO, McDermott M, Sinha AA, Timmerman L, Steinman L, McDevitt HO: A molecular basis for MHC class IIassociated autoimmunity. Science 240:1003-1009, 1988

Journal of Clinical Immunology, VoL 10, No. 6 (November Supplement 1990)

38S

STEINMAN

3. Steinman L, Rosenbaum JT, Sriram S, McDevitt HO: In vivo effects of antibodies to immune response gene products: Prevention of experimental allergic encephalitis. Proc Natl Acad Sci USA 78:7111-7114, 1981 4. Sriram S, Steinman L: Anti I-A antibody suppresses active encephalomyelitis: Treatment model for diseases linked to IR genes. J Exp Meal 158:1362-1367, 1983 5. Jonker M, van Lambalgen R, Mitchell DJ, Durham SK, Steinman L: Successful treatment of EAE in rhesus monkeys with MHC class II specific monoclonal antibodies. J Autoimmun 1(5):399-414, 1988 6. Zamvil SS, Mitchell DJ, Moore AC, Kitamura K, Steinman L, Rothbard JB: T-celt epitope of the autoantigen myelin basic protein that induces encephalomyelitis. Nature (London) 324:258-260, 1986 7. Zamvil SS, Mitchell DJ, Powell MB, Sakai K, Rothbard JB, Steinman L: Multiple discrete encephaIitogenic epitopes of the autoantigen myelin basic protein include a determinant for I-E class II-restricted T cells. J Exp Med 168:1181-1186, 1988 8. Sakai K, Sinha A, Mitchell DJ, Zamvil SS, McDevitt HO, Rothbard JB, Steinman L: Involvement of distinct T cell receptors in the autoimmune encephalitogenic response to nested epitopes of myelin basic protein. Proc Natl Acad Sci USA 85:8608-8612, 1988 9. Acha-Orbea H, Mitchell DJ, Timmermann L, Wraith DC, Tausch GS, Waldor MK, Zamvil SS, McDevitt HO, Steinman L: Limited heterogeneity of T cell receptors from lymphocytes mediating autoimmune encephalomyelitis allows specific immune intervention. Cell 54:263-273, t988 10. Urban JL, Kumar V, Kono DH, Gomez C, Horvath SJ, Clayton J, Ando DG, Sercarz EE, Hood L: Restricted use of T cell receptor V genes in murine autoimmune encephalomyelitis raises possibilities for antibody therapy. Cell 54:577-592, 1988 11. Vanderbark AA, Hashim G, Offner H: Immunizationwith a synthetic T-cell receptor V-region peptide protects against experimental encephalomyelitis. Nature 341:541-544, 1989 12. Howell MD, Winters ST, Olee T, Powell HC, Carlo DJ, Brostoff SW: Vaccination against experimental allergic encephalomyelitis with T cell receptor peptides. Science 246:668-670, 1989 13. Wraith DC, Smilek DE, Mitchell DJ, Steinman L, McDevitt HO: Antigen recognition in autoimmune encephalomyelitis and the potential for peptide mediated immunotherapy. Cell 59:247-255, 1989 14. Sakai K, Zamvil SS, Mitchell DJ, Hodgkinson S, Rothbard JB, Steinman L: Prevention of experimental encephalomyelitis with peptides that block interaction of T cells with major histocompatibility complex proteins. Proc Natl Acad Sci USA 86:9470-9474, 1989 15. Urban JL, Horvath SJ, Hood L: Autoimmune T cells: Immune recognition of normal and variant peptide epitopes and peptide-based therapy. Cell 59:957-271, 1989 16. Todd JA, Bell JI, McDevitt HO: A molecular basis for genetic susceptibility to insulin dependent diabetes mellitus. Trends Genet 4:129-134, 1988 17. Sinha AA, Brautbar C, Szafer F, Friedmann A, Tzfoni E, Todd JA, Steinman L, McDevitt HO: A newly characterized HLA-DQ beta allele associated with pemphigus vulgaris. Science 239:1026-1029, 1988

18. Schaff S, Friedmann A, Brautbar C, Szafer F, Steinman L, Horn G, Gyllenstein U, Erlich HA: HLA class II allelic variation and susceptibility in pemphigus vulgaris. Proc Natl Acad Sci USA 85:3504-3508, 1988 19. Oksenberg JR, Sherritt M, Begovich AB, Erlich HA, Bernard CCA, Cavalli-Sforza LL, Steinman L: T cell receptor V alpha and C alpha alleles associated with multiple sclerosis and myasthenia gravis. Proc Natl Acad Sci USA 86:988-992, 1989 20. Oksenberg JR, Stuart S, Begovich A, Bell R, Erlich HA, Steinman L, Bernard CCA: Limited heterogeneity of T cell receptor V alpha transcripts in brains of multiple sclerosis patients. Nature 345:344-346, 1990 21. Brocke S, Brautbar C, Steinman L, Abramsky O, Rothbard J, Neumann D, Fuchs S, Mozes E: In vitro proliferative responses and antibody titres specific to human acetylcholine receptor synthetic peptides in patients with myasthenia gravis and relation to HLA class II genes. J Clin Invest 82:1894-1900, 1988

DISCUSSION

Dr. Yolken: In your interesting studies on multiple sclerosis and the V alpha, I noticed that the factor you were looking at was found also in the controls, although there was a highly statistical difference between the frequency in the controls and that in the cases. What is your theory about what is happening? Is there an environmental factor which allows for the genetics to be played out and therefore puts those people at risk or is there a second genetic factor that you are not picking up or a genetic rearrangement? Dr. Steinman: In any disease in which a certain percentage of healthy persons carries the susceptibility gene (or susceptibility polymorphism) and a larger percentage of people with the disease has that gene, there is usually more than a single gene involved. In an autoimmune process we know that at least T cells and MHC are involved so that maybe if one did not have the appropriate TCR, or if the particular MHC did not allow presentation, one would not get sick. But we know the human immune system is a lot more complicated than that. For example, for pemphigus and myasthenia the B-cell repertoire also has to be able to make the autoantibody. So, because it is a stochastic interaction, having a single bad gene may not lead to the disease. Is this a response to an environmental pathogenic material? In my opinion, probably yes. I believe we will find that many autoimmune diseases are triggered by routine immune responses to pathogens. The idea of molecular mimicry appeals to me. If one has the wrong TCR genotype and the wrong MHC genotype and acquires a streptococcal

Journal o f Clinical Immunology, VoL 10, No. 6 (November Supplement 1990)

MONOCLONAL ANTIBODIES FOR AUTOIMMUNE DISEASE

or a Yersinia infection and happens to respond to fragments of those pathogens that mimic self, then the disease develops. I find that concept very appealing because there is no need to assume anything special about autoimmune disease; it is simply an extension of normal immunity to infectious pathogens. However, in the diseases I work with, we do not have any good candidates for the pathogen. There are some very good candidates that can be brought in for questioning in some of the rheumatologic diseases but whether they will turn out to be the culprits that cause the disease we do not know. Dr. Ochs: Nonselected gamma globulin has been tried clinically, and successfully to some extent, to treat patients with autoimmune diseases and your data concerning the peptides may explain w h y it works. Could you speculate on the possibility that a large pool of immune globulin might contain specific antibodies to these peptides which would modulate that procedure in the autoimmune diseases? Dr. Steinman: I believe it is certainly possible. Dr. Levinson: We hear a lot about the use of ihtravenous immune globulin (IVIG) in autoimmune disease. Is it theoretically possible that when one administers immune globulin at high levels, it is competing for self-peptides on MHC-positive antigen-presenting cells? Dr. Steinman." It is possible. Paul Allen showed that normal hemoglobin can lie in the MHC-binding cleft. But the question is, H o w does one go from speculation to the design of an experiment? I do not know. If we could design such an experiment, I would not be surprised if we were to show that that is the way IVIG works. Dr. Desai: My question concerns two challenging patients of mine. One is about 70 years old and has chronic lymphocytic leukemia with counts of about 30,000 and benign disease. She has not been treated for the last 2 years. About 6 months ago she developed an increasing amount of difficulty in swallowing and a neurology colleague diagnosed amyotrophic lateral sclerosis. To the best of my knowledge, no treatment is available for this disease. Could anti-Tac or peptide be useful for this

39S

case? The second patient is a 50-year-old female who came in with lymphocytic leukemia, a total white-cell count of about 13,000 to 20,000, and total alopecia, a really remarkable picture of hair loss throughout the body over a period of just a few weeks. The patient was treated with isoprinosone and did not do well. She was then treated with interferon and still is not doing well. She has skin lesions and she has proven T-cell lymphoma. I wonder whether either of the two treatment modalities you described could be useful? Dr. Steinman: What kind of tymphoma does she have? Dr. Desai: T-cell lymphoma. She has cutaneous lesions. Dr. Steinman: I have wanted to give anti-CD4 to people with autoimmune disease for the last 5 years. We have shown that it works very well in the EAE model and we requested permission from the FDA to use chimeric anti-CD4 but the FDA refused, saying multiple sclerosis is too "gentle" a disease to try that kind of therapy. They did give us permission to try it in cutaneous T-cell lymphoma. Ron L e v y led the project and it worked quite dramatically well; there was a marked reduction in the erythroderma and in the tumor bulk. Among the seven patients who received the chimeric anti-CD4 (made by Becton Dickinson), a few showed a weak antivariable region response. The immunosuppression one would expect to see with an anti-CD4 in terms of reduction in the mixed lymphocyte reaction activity was observed. One patient had cutaneous T-cell lymphoma and rheumatoid arthritis and her rheumatoid arthritis cleared up during the course of the treatment, so I feel it would be worthwhile to pursue that modality. These patients with mycosis fungoides were really sick, but after treatment they walked out of the hospital, they dressed themselves, and some of them resumed playing golf. The unfortunate patients I wanted to treat are usually wheelchair bound and unable to feed and dress themselves. As a neurologist, I was very disappointed that we were not permitted to treat them.

Journal of Clinical Immunology, VoL 10, No. 6 (November Supplement 1990)