Gajdusek and his group, meanwhile, had also been trying to produce an animal model for lytico-bodig and had been maintaining a number of macaques on a low-calcium, high-aluminum diet. The monkeys developed no clinical symptoms in the four years of the trial, but autopsies showed many neurofibrillary tangles, as well as degenerative changes in the motor neurons, throughout the neuraxis. These changes seemed to resemble those of lytico-bodig or the presymptomatic changes described by Anderson and Chen, and it was speculated that a longer period of calcium deficiency, or higher doses of toxic metals, might have led to overt clinical disease. And though Gajdusek had told John in 1983 that he thought the lytico-bodig was dying out in Guam, he has continued to investigate it in Irian Jaya, where in 1993 he found it still had a remarkably high incidence. He and his colleagues continue to see aluminum neurotoxicity as the cause of lytico-bodig and indeed of a wide range of other conditions.
While Spencer, for his part, was greatly encouraged by his own success in inducing neurological disorders in primates with BMAA, he soon developed reservations. The disorders shown by his monkeys were dose-related, came on promptly, and were acute and nonprogressive (they resembled, in this way, the neu-rocycadism of cattle); whereas human lytico-bodig, it was abundantly clear, had a very long latency or incubation period, but once it had become symptomatic, was almost invariably progressive. Was it possible, Spencer speculated, that another factor was involved besides the BMAA, which might not predispose to overt disorder for many years? Slow viruses had been described by Gajdusek; could there not be, analogously, a slow toxin? Spencer did not have any clear idea, at this stage, of how such a toxin might work, or any way of validating the concept.
Though Gajdusek might have been expected to be sympathetic to the idea of a slow toxin, he argued passionately against it in a sternly titled paper, ‘Cycad Toxicity Not the Cause of High-Incidence ALS⁄Parkinsonism-Dementia on Guam, Kii Peninsula of Japan, or in West New Guinea,’ asserting that such a hypothesis was, first, redundant; second, without precedent; third, without support; and fourth, impossible:
No neurotoxin has been demonstrated to give rise to fatal central nervous system disease, neurological signs and symptoms of which first start to be detectable years after exposure to the neurotoxin has ceased. In fact, we have no example of any toxin producing progressive damage to any organ years after last exposure to the substance…Only hypersensitivity disorders, slow infections, and genetically-timed disorders have given rise to this pattern of long delay.
Spencer, undeterred, saw Gajdusek’s words as a challenge (indeed he has cited them in several of his own papers), and continued to see his task as the search for a new kind of toxin, a new kind of toxic mechanism, hitherto unrecognized in medicine. A great deal of attention was focused, in the sixties and seventies, on carcinogenesis, the appearance of cancers, in some cases, years after an initial exposure to the carcinogen, whether radioactivity, a toxin, or a virus. It had been established, in Kur-land’s original cycad conferences, what a potent carcinogen cycasin was, capable of inducing liver cancers and colon and kidney malformations. It had been observed, moreover, that if infant rats were fed cycasin-high diets, the still-dividing Purkinje cells of the cerebellum might develop bizarre multinucleated forms and ectopic ‘nests,’ and such findings had also been reported, on occasion, in cases of human lytico-bodig.
What then might be the effect of cycasin, Spencer wondered, on adult nerve cells, which are no longer capable of dividing?
He and Glen Kisby have postulated recently that cycasin (or its component, MAM, or methazoxymethanol) may be able to form stable compounds with the DNA in nerve cells (such adduct formation is believed to underlie the overt carcinogenic and teratogenic effects of cycasin elsewhere in the body). This aberrant DNA in the nerve cells, he thinks, could lead to subtly but persistently altered metabolic functions, the nerve cells finally becoming oversensitive to their own neurotransmitters, their own glutamate, so that this itself could act as an excitotoxin. No external agent would be needed to provoke a neurological disaster at this point, for in this pathologically sensitized state, even normal neural functioning would now overexcite neurotransmitter receptor cells and push them toward their own destruction.
The notion of such a gene toxin is not as outlandish as it seemed a decade ago, and Spencer and Kisby have now observed DNA changes in tissue cultures of cells exposed to cycasin which suggest that such a mechanism may be at work in lytico-bodig. Such a gene toxin would actually alter the genetic character of the nerve cells it affected, producing, in effect, a genetically-based form of hypersensitivity disorder.
Now that Spencer was pondering the possible effects of cycasin on adult nerve cells, he had new analyses made of traditionally prepared cycad flours and found (contrary to what John had found earlier) that the Guam samples, even though low in BMAA, contained substantial amounts of cycasin. The highest levels of cycasin, indeed, were found in samples from villages with the greatest prevalence of lytico-bodig, lending strong circumstantial support to the hypothesis of cycasin toxicity.[72]
John is a very vivid storyteller, and as he told me this story – a story not just of a scientific odyssey, but of his own most passionate hopes and disappointments – he seemed to relive it with almost unbearable intensity. He had enjoyed a cordial relationship with Kurland and Gajdusek, he thought, and a passionate one with Spencer – but when, in 1990, he gave up on the cycad hypothesis (as four years earlier he had given up on the mineral hypothesis), a sense of intense isolation gripped him; he felt he was out in the cold, seen as an apostate by them all. In the early 1990s, he toyed with a viral hypothesis (this was very much in his mind when we first met, in 1993). But as a primary physician, a general practitioner, living and working amidst the whole affected community in Umatac, he has been forced to think in terms of the entire families or clans with lytico-bodig under his care – no external cause alone, it seemed clear, could adequately account for such a pattern of distribution. Had a genetic theory been thrown out prematurely? Much had changed since Kurland and Mulder first considered, then rejected, this in the 1950s. The classic Mendelian patterns of inheritance had now been joined by concepts of complex inheritance involving the presence of multiple genetic abnormalities and their interactions with each other and with environmental factors. Further, it was now possible to directly examine the genetic material with molecular biology, using technologies and concepts not available to the early investigators.
Working with Verena Keck, an anthropologist, John started to collect pedigrees of every patient he had seen – pedigrees of unprecedented accuracy and detail, including medical histories going back fifty years. The more pedigrees he obtained, the more he became convinced that there had to be some genetic predisposition, or perhaps several predispositions – for it looked as though the lytico and the bodig had different patterns in different families. Sometimes one saw a family in whom the affected members had only the lytico, sometimes a family in which the clinical expression was always bodig, and sometimes, more rarely, a family with both. The similarity of the pathological pictures in lytico and bodig, he started to feel, might have been misleading them all; genealogically, they seemed to be two separate diseases.