Autoimmune Folate Deficiency and the Rise and Fall of "Horror Autotoxicus"
http://www.100md.com
《新英格兰医药杂志》
In 1904, a little more than a century ago, Julius Donath, a senior physician working in the Vienna Merchants' Hospital, and Karl Landsteiner, the chair of the pathology department at the University of Vienna and one of the founders of immunochemistry, reported the first known autoimmune disease in humans, paroxysmal cold hemoglobinuria. They showed that an antibody in this disease reacted in the cold with the patient's own erythrocytes. By definition, the Donath–Landsteiner antibody was an autoantibody. This revolutionary discovery flew in the face of "horror autotoxicus," a principle that Paul Ehrlich had expounded after failing to find autoantibodies in goats that had been immunized with their own red cells.
By all rights, the revelation of the mechanism of paroxysmal cold hemoglobinuria should have swept away any doubts about autoimmunization as a mechanism of disease and begun a golden era in medical research. Yet the very opposite happened. For more than 50 years after the Donath–Landsteiner report, the field of autoimmunization dwelled in a dark age in which the belief in the incontrovertibility of horror autotoxicus squelched any proposition that autoantibodies might cause a disease. It is hard to believe today that the autoantibodies on the red cells in autoimmune hemolytic anemia were once called "erythrocyte-coating substances" and that antibodies against nuclear antigens in systemic lupus erythematosus were termed "antinuclear factors." Such verbal fig leaves covered an embarrassing misperception of the obvious. Even so, there were some who grudgingly acknowledged that an autoimmune disease might exist, but who hedged by referring to it rhetorically as "so-called autoimmune disease."
All that is over now, and the study of autoimmunization has at last entered its golden age. Today, autoimmunization lurks behind every pillar, ready to strike from cerebellum to calcaneum and to play its part in disorders ranging from diabetes to childhood neurodegenerative diseases. With each new disclosure, however, the puzzle deepens. Ehrlich, after all, was right: the body does abhor autoimmunization and deploys elaborate mechanisms to avoid it. How these mechanisms break down, thereby permitting B cells to produce pathogenic autoantibodies and T cells to turn against their owner, is a central problem in immunology.
In this issue of the Journal, Ramaekers and colleagues (pages 1985–1991) report the formation, in children with an unusual neurodegenerative disorder, of autoantibodies against folate receptors that block the transport of folate (in the form of its active metabolite 5-methyltetrahydrofolate) through the choroid plexus into the central nervous system. The choroid plexus is rich in folate receptors, and the blockade by the autoantibodies deprives the baby's developing brain of an essential vitamin. The folate level in the child's cerebrospinal fluid is low, whereas it is normal in blood — probably because in other tissues, transporters like the reduced folate carrier bypass the blockade of folate receptors. The result of the autoantibody-mediated blockade is a disorder called infantile-onset cerebral folate deficiency syndrome. Its main features are irritability, psychomotor retardation, ataxia, dyskinesias, and pyramidal tract signs in the legs. Autoantibodies against placental folate receptors have also been found in women in whom a current or previous pregnancy has been complicated by a neural-tube defect.
Antibodies against folate receptors are representative of a group of antireceptor autoantibodies, some of which block binding of the native ligand to the receptor and others of which are agonists that mimic the action of the ligand on the receptor. In myasthenia gravis, autoantibodies against the acetylcholine receptor block neuromuscular transmission at the motor endplate. In Graves' disease, by contrast, autoantibodies against thyrotropin receptors stimulate thyroid follicular cells to secrete excessive amounts of thyroxine. Autoantibodies that decrease cardiac adenylate cyclase activity by blocking the cardiac 1-adrenoreceptor have been found in idiopathic dilated cardiomyopathy and in the cardiomyopathy of Chagas' disease. Blocking autoantibodies that impair the binding of vitamin B12 to intrinsic factor occur in pernicious anemia, which sometimes coexists with two other autoimmune disorders, type 1 diabetes and hypothyroidism. There are blocking autoantibodies against the insulin receptor in some cases of hypoglycemia and insulin resistance, and agonistic autoantibodies against the high-affinity IgE receptor have been found in chronic urticaria. Autoantibodies against the angiotensin I receptor in the serum of women with preeclampsia cause vascular smooth-muscle cells to produce the coagulant protein tissue factor, the probable cause of placental infarction in preeclampsia.
All these antireceptor autoantibodies are of special interest because two elements are involved: the receptor and its ligand — a problematic duality. Does the receptor or the ligand incite the autoantibodies? It is not easy to show experimentally how autoantibodies arise against a receptor as ubiquitous as the folate receptor. Molecular mimicry is one possibility. Ramaekers et al. speculate that structurally similar soluble folate-binding proteins in human or cow's milk may be able to provoke anti–folate receptor autoantibodies (see diagram, Panel A). Another example of possible molecular mimicry is the cardiomyopathy of Chagas' disease, in which a ribosomal protein of Trypanosoma cruzi has a structural resemblance to the 1-adrenoreceptor.
Possible Routes to the Induction of Antireceptor Autoantibodies.
Each hypothesis would require a genetic predisposition to autoimmunization and an immunogenetic permissiveness that would allow the mechanism to proceed. Soluble folate-binding peptides in milk (Panel A) may induce antibodies that cross-react with the child's folate receptors. Cell-bound thyrotropin receptors (Panel B) may degrade, and the resulting peptide fragments may enter the blood and stimulate immune cells to produce antibodies that bind to native thyrotropin receptors. An antibody against a protein unrelated to the receptor (Panel C) may have a binding region with a configuration that allows it to bind to the receptor. Or an antibody against the ligand (Panel D) may incite the production of an anti-idiotypic antibody whose binding site is the same shape as the ligand.
Structural features of the receptor itself may be important. For example, the thyrotropin receptor is normally cleaved and the resultant fragments enter the blood, where they might, as blemished remnants, come into contact with immune cells (see diagram, Panel B).
The theory that the ligand provokes antireceptor autoantibodies is attractive, because receptors are not always specific for a single ligand, making it possible that a third party is involved. Moreover, some receptors for physiologic ligands are gateways for viral entry. An antibody that has no relation to the ligand but whose three-dimensional shape coincidentally resembles the ligand's contours or those of a virus that uses the receptor could fit into the receptor without being an antireceptor antibody in a formal sense (see diagram, Panel C).
Another possibility stems from idiotypes, the potentially immunogenic configurations of the variable region of an antibody, which can provoke anti-idiotypes that resemble the ligand (or virus [see diagram, Panel D]). There is experimental evidence that anti-idiotypes can behave like antireceptor antibodies, but digging relevant anti-idiotypes from serum without knowledge of the idiotype is a formidable task.
There is a vast literature on experimental autoimmune diseases of the type that is induced in inbred rodents by immunization with potent adjuvants. These models have been instructive, but there is an as yet unbridgeable chasm between a mouse with experimental autoimmune encephalitis and a patient with multiple sclerosis. There is an even larger literature on how the immune system of a laboratory animal responds to foreign antigens but usually not to its own. This, too, is instructive, and it has provided immunologists with many talking points. But clinicians still have little to say when a patient with autoimmune hemolytic anemia asks about the cause of the disease. Nevertheless, there is reason for hope that we will get to the bottom of the problem of autoimmune diseases within the coming decades, because clinical investigators now have new tools and interesting ideas about these diseases. Potent and costly, yet clinically useful, new drugs that ]reduce the inflammation induced by autoimmunization are now available, but they do no more than suppress the disease. Cures will come only when the causes of autoimmunization are finally uprooted.(Robert S. Schwartz, M.D.)
By all rights, the revelation of the mechanism of paroxysmal cold hemoglobinuria should have swept away any doubts about autoimmunization as a mechanism of disease and begun a golden era in medical research. Yet the very opposite happened. For more than 50 years after the Donath–Landsteiner report, the field of autoimmunization dwelled in a dark age in which the belief in the incontrovertibility of horror autotoxicus squelched any proposition that autoantibodies might cause a disease. It is hard to believe today that the autoantibodies on the red cells in autoimmune hemolytic anemia were once called "erythrocyte-coating substances" and that antibodies against nuclear antigens in systemic lupus erythematosus were termed "antinuclear factors." Such verbal fig leaves covered an embarrassing misperception of the obvious. Even so, there were some who grudgingly acknowledged that an autoimmune disease might exist, but who hedged by referring to it rhetorically as "so-called autoimmune disease."
All that is over now, and the study of autoimmunization has at last entered its golden age. Today, autoimmunization lurks behind every pillar, ready to strike from cerebellum to calcaneum and to play its part in disorders ranging from diabetes to childhood neurodegenerative diseases. With each new disclosure, however, the puzzle deepens. Ehrlich, after all, was right: the body does abhor autoimmunization and deploys elaborate mechanisms to avoid it. How these mechanisms break down, thereby permitting B cells to produce pathogenic autoantibodies and T cells to turn against their owner, is a central problem in immunology.
In this issue of the Journal, Ramaekers and colleagues (pages 1985–1991) report the formation, in children with an unusual neurodegenerative disorder, of autoantibodies against folate receptors that block the transport of folate (in the form of its active metabolite 5-methyltetrahydrofolate) through the choroid plexus into the central nervous system. The choroid plexus is rich in folate receptors, and the blockade by the autoantibodies deprives the baby's developing brain of an essential vitamin. The folate level in the child's cerebrospinal fluid is low, whereas it is normal in blood — probably because in other tissues, transporters like the reduced folate carrier bypass the blockade of folate receptors. The result of the autoantibody-mediated blockade is a disorder called infantile-onset cerebral folate deficiency syndrome. Its main features are irritability, psychomotor retardation, ataxia, dyskinesias, and pyramidal tract signs in the legs. Autoantibodies against placental folate receptors have also been found in women in whom a current or previous pregnancy has been complicated by a neural-tube defect.
Antibodies against folate receptors are representative of a group of antireceptor autoantibodies, some of which block binding of the native ligand to the receptor and others of which are agonists that mimic the action of the ligand on the receptor. In myasthenia gravis, autoantibodies against the acetylcholine receptor block neuromuscular transmission at the motor endplate. In Graves' disease, by contrast, autoantibodies against thyrotropin receptors stimulate thyroid follicular cells to secrete excessive amounts of thyroxine. Autoantibodies that decrease cardiac adenylate cyclase activity by blocking the cardiac 1-adrenoreceptor have been found in idiopathic dilated cardiomyopathy and in the cardiomyopathy of Chagas' disease. Blocking autoantibodies that impair the binding of vitamin B12 to intrinsic factor occur in pernicious anemia, which sometimes coexists with two other autoimmune disorders, type 1 diabetes and hypothyroidism. There are blocking autoantibodies against the insulin receptor in some cases of hypoglycemia and insulin resistance, and agonistic autoantibodies against the high-affinity IgE receptor have been found in chronic urticaria. Autoantibodies against the angiotensin I receptor in the serum of women with preeclampsia cause vascular smooth-muscle cells to produce the coagulant protein tissue factor, the probable cause of placental infarction in preeclampsia.
All these antireceptor autoantibodies are of special interest because two elements are involved: the receptor and its ligand — a problematic duality. Does the receptor or the ligand incite the autoantibodies? It is not easy to show experimentally how autoantibodies arise against a receptor as ubiquitous as the folate receptor. Molecular mimicry is one possibility. Ramaekers et al. speculate that structurally similar soluble folate-binding proteins in human or cow's milk may be able to provoke anti–folate receptor autoantibodies (see diagram, Panel A). Another example of possible molecular mimicry is the cardiomyopathy of Chagas' disease, in which a ribosomal protein of Trypanosoma cruzi has a structural resemblance to the 1-adrenoreceptor.
Possible Routes to the Induction of Antireceptor Autoantibodies.
Each hypothesis would require a genetic predisposition to autoimmunization and an immunogenetic permissiveness that would allow the mechanism to proceed. Soluble folate-binding peptides in milk (Panel A) may induce antibodies that cross-react with the child's folate receptors. Cell-bound thyrotropin receptors (Panel B) may degrade, and the resulting peptide fragments may enter the blood and stimulate immune cells to produce antibodies that bind to native thyrotropin receptors. An antibody against a protein unrelated to the receptor (Panel C) may have a binding region with a configuration that allows it to bind to the receptor. Or an antibody against the ligand (Panel D) may incite the production of an anti-idiotypic antibody whose binding site is the same shape as the ligand.
Structural features of the receptor itself may be important. For example, the thyrotropin receptor is normally cleaved and the resultant fragments enter the blood, where they might, as blemished remnants, come into contact with immune cells (see diagram, Panel B).
The theory that the ligand provokes antireceptor autoantibodies is attractive, because receptors are not always specific for a single ligand, making it possible that a third party is involved. Moreover, some receptors for physiologic ligands are gateways for viral entry. An antibody that has no relation to the ligand but whose three-dimensional shape coincidentally resembles the ligand's contours or those of a virus that uses the receptor could fit into the receptor without being an antireceptor antibody in a formal sense (see diagram, Panel C).
Another possibility stems from idiotypes, the potentially immunogenic configurations of the variable region of an antibody, which can provoke anti-idiotypes that resemble the ligand (or virus [see diagram, Panel D]). There is experimental evidence that anti-idiotypes can behave like antireceptor antibodies, but digging relevant anti-idiotypes from serum without knowledge of the idiotype is a formidable task.
There is a vast literature on experimental autoimmune diseases of the type that is induced in inbred rodents by immunization with potent adjuvants. These models have been instructive, but there is an as yet unbridgeable chasm between a mouse with experimental autoimmune encephalitis and a patient with multiple sclerosis. There is an even larger literature on how the immune system of a laboratory animal responds to foreign antigens but usually not to its own. This, too, is instructive, and it has provided immunologists with many talking points. But clinicians still have little to say when a patient with autoimmune hemolytic anemia asks about the cause of the disease. Nevertheless, there is reason for hope that we will get to the bottom of the problem of autoimmune diseases within the coming decades, because clinical investigators now have new tools and interesting ideas about these diseases. Potent and costly, yet clinically useful, new drugs that ]reduce the inflammation induced by autoimmunization are now available, but they do no more than suppress the disease. Cures will come only when the causes of autoimmunization are finally uprooted.(Robert S. Schwartz, M.D.)