Genentech

A striking fact of Genentech's earliest years is that, for a while, it had almost no laboratory of its own. It had a concept and very little money, and money for benches and freezers and skilled hands was exactly what it lacked. So the company did something unusual: it contracted the science out, paying university and institute laboratories to run its experiments while it held the commercial rights to whatever they produced. The work that would make Genentech's name happened, at first, in facilities the company did not own.

The first target was chosen for its symbolic clarity as much as its ambition. In 1977 the team set out to make a bacterium produce somatostatin, a human brain hormone — not because there was a market for it, but because succeeding would prove that the trick worked at all, that a microbe could be made to build a human protein to order. The work was carried out largely at City of Hope, a research hospital near Los Angeles, in collaboration with chemists there. When the bacteria yielded the hormone, it was the first demonstration that recombinant DNA could manufacture a human substance.

Insulin was the obvious prize. Diabetics depended on it, and the world's supply came from the pancreases of slaughtered cattle and pigs — a system that was workable but limited, occasionally allergenic, and clearly finite against rising demand. A human insulin made by bacteria would be purer, effectively unlimited, and a genuine product. It also carried a scientific weight the company understood well: whoever synthesized human insulin first would have proven that this technique belonged in medicine, not just in journals.

Genentech was not alone. A group at Harvard, and another at UCSF, were pursuing the same goal, and the competition was intense enough that Hughes tells this stretch as a race, with the anxieties and secrecy that races produce. Genentech again leaned on the City of Hope chemists to assemble the genes and on its bacterial expertise to express them. In the late summer of 1978 the team succeeded in getting bacteria to produce the two chains of human insulin, which could then be joined into the functioning hormone.

The 1980 offering is remembered for its numbers, but the more lasting thing it did was change what a number could mean. Genentech's shares were not priced against earnings, because there were none. They were priced against a promise — against insulin licensed but not yet sold, against proteins proven in principle, against the plausible future of an entire technique. The run from $35 to $89 was the market discovering that it was willing to buy a scientific expectation as if it were a balance sheet. That willingness did not disappear when the trading stopped; it became a template.

What Genentech established, more than any single molecule, was a way of financing science. A company could go public on the strength of its research pipeline and its patents rather than its profits, and investors would fund years of expensive experiment in exchange for a stake in what might come. The biotechnology industry that followed — hundreds of firms, most of them unprofitable for a decade or more — ran on exactly this arrangement. The IPO of a company without products became a recognized, respectable event because Genentech had made it one first.

Genentech had begun in 1976 as two men and a thousand dollars, betting that a technique confined to research journals could be pointed at something people would pay for. Four years later, on a single October morning, the market agreed in the most extravagant terms it had, running the stock from $35 to $89 before lunch. The company still had no drug of its own on any shelf and had not turned a profit. What it had was somatostatin, human insulin licensed to Lilly, and a reputation for doing what others called premature.