We Failed Steve Jobs

Only five in 1 million people suffer from Jobs’s form of cancer. Evens Lee / Color China Photo-AP

In Oct. 5, the night that Steve Jobs died, I ascended 30,000 feet into the thin air above New York on a flight to California. On my lap was a stash of scientific papers. I was reading and taking notes—where else?—on an iPad.

Jobs’s death—like a generational Rorschach test—had provoked complex reactions within each of us. There was grief in abundance, of course, admixed with a sense of loss, with desolation and nostalgia. Outside the Apple store in SoHo, New York, that evening, there were bouquets of white gerberas and red roses. Someone had left a bushel of apples by the doorstep and a sign that read “I-miss ...”

I missed Jobs, too—but I also felt a personal embarrassment in his death. I am an oncologist and a cancer researcher. I felt as if my profession, my discipline, and my generation had let him down. Steve Jobs had promised—and then delivered—life-altering technologies. Had we, in all honesty, given him any such life-altering technologies back?

I ask the question in all earnestness. Jobs’s life ended because of a form of pancreatic cancer called pancreatic neuroendocrine tumor, or PNET. These tumors are fleetingly rare: about five in every million men and women are diagnosed with PNETs each year. Deciphering the biology of rare cancers is often challenging. But the past five years have revealed extraordinary insights into the biology of some rare cancers—and PNETs, coincidentally enough, have led part of that charge. By comparing several such tumors, scientists are beginning to understand the biology of these peculiar tumors.

But understanding biology is an abstract activity. Steve Jobs needed more than “biology.” He needed medicines. And despite our efforts, we were unable to transform our knowledge about PNETs into medical realities during his lifetime. The question is, are we ready to achieve this transformation sometime in the near future?

Let’s take PNETs as a case in point. In 2008 a team of scientists from Johns Hopkins University set out to document all the gene mutations in PNETs—creating a systematic genetic “anatomy” of these tumors. Cancer, of course, is ultimately a disease of mutations in genes. Human cells possess about 23,000 genes in total. In cancer cells some of these genes are changed—mutated—and begin to function abnormally.

Many of the genes that are mutated in various cancers, predictably, control cellular growth. Genes regulate the growth of cells like invisible puppeteers tugging and pushing opposing strings behind curtains. There are genes that command a cell to grow and those that tell a cell to stop growing. Cancer occurs when these growth-control genes are mutated, resulting in the dysregulated growth of a cell.

But the genes most frequently mutated in PNETs are odd. They don’t seem to control growth directly; rather they seem to affect the way cells regulate genes. Take a moment to understand this by considering normal development. A cell in the retina possesses the same 23,000 genes as a cell in the skin, yet these two cells barely resemble each other in shape, size, or behavior. How does a cell, then, “know” how to become a retinal cell versus a skin cell? How can the same set of 23,000 genes be used to specify such radically diverse behaviors, functions, and forms?

Cancer Budget Graphic

Part of the answer lies in the way genes are controlled, or regulated. Although a skin cell and a retinal cell inherit the same set of 23,000 genes, a skin cell activates or suppresses a unique subset of the total—say 5,000 of the 23,000—while a retinal cell activates another subset. It’s like an elaborate mix-and-match game: each cell dips into the same box of genes and chooses a unique spectrum of genes for itself, thereby attaining its form, function, and behavior.

To use an analogy that Jobs might have used: every human cell contains the same “hardware” of genes. But every individual cell type activates a particular “software”—a program (involving a particular combination of genes) that is unique unto itself to achieve its particular function.

But the answer raises a question: how does the skin cell know how to activate such software? In part through master regulatory genes that accomplish this task. These master-control genes exercise exquisite control on the growth, shape, size, and identity of a cell. They activate entire programs of gene expression; they toggle hundreds of molecular switches to turn “on” and turn “off” programs.

There are many such families of master-control genes, and one such family acts by modifying DNA—the stuff that all genes are made of. And PNETs appear to possess frequent mutations in these DNA-modifying genes. The phenomenon is not unique to PNETs. In my own laboratory we have discovered that leukemias and other blood disorders also possess frequent changes in such DNA-modifying genes. Others have found mutations in DNA-modifying genes in lymphomas and brain tumors, and in colon and stomach cancers.

There is, in short, a novel principle of cancer unfolding here: that DNA-modifying genes can be mutated in certain tumors. But what connects these DNA-modifying genes to the ultimate growth behavior of a cancer cell? Might these genes become targets for new drugs? Might one such medicine be used to treat—or even cure—PNETs?

Indeed, DNA-modifying genes aren’t the only new and unusual targets unfolding as we learn more about cancer. There are genes that affect the way cancer cells use glucose or other building blocks, such as amino acids, during their metabolism.

This knowledge—gleaned over the past five years—should have accelerated an effort to find medicines that affect these new pathways that seem to control cancer. And indeed, to an extent, it has. Many of these new pathways have become reasonable targets for anti-cancer drugs. Medicines that attack a specific family of genes termed “kinases”—genes dysregulated in blood and lung cancers and melanoma—have made their way into human use. Nearly all of these originated with the discovery of specific gene mutations in particular variants of cancer.

Cancer Rate Statistics Graphic

But there are vast gaps in knowledge still, and even larger chasms in drug development. Targeting an errant gene in a cancer cell might sound like a simple, well-defined task, but in fact it is mind-bogglingly complex. Cancer cells arise out of normal cells—and they resemble normal cells so closely that it can be difficult, at times, to tell them apart at a genetic level (of the 23,000 genes in a normal cell, a cancer cell may share 22,962 and have only 38 that are altered).

I wonder whether Jobs himself might have enjoyed the strange challenge of creating anti-cancer drugs: it is, after all, the ultimate design problem. Molecules have to be made to “fit” exactly into unique clefts and pockets of a cell’s machinery in order to logjam malignant growth. Extraneous bits and pieces have to be shaved off so that the drug can bind to its intended target with a neat, satisfying click. The human testing and clinical trials that follow drug discovery present peculiar operational challenges. And there’s production and safety monitoring that come after. The popular press has adopted the term “designer drugs” for the new generation of molecules that can target cancer cells with exquisite specificity. I like to think that the ultimate designer of our generation might have had something to add to this most profound frontier of design.

Certainly he could have added a plea for adequate resources. The National Cancer Institute (NCI) is the nation’s preeminent institution tasked with leading cancer scientists toward new means to prevent, treat, and cure cancers. Its annual budget of about $5 billion is stagnating and threatened. The Food and Drug Administration, charged with ensuring that new cancer medicines are brought effectively and safely to the public, operates on a budget of about $4 billion. If these amounts sound impressive, consider the fact that in 2008 the United States was spending about $12 billion every month on conflict in the Middle East—more than the annual NCI and FDA budgets combined.

If we are truly committed to creating medicines to prevent and treat cancer, we seem to be doing a rather lax job in funding this effort. We cannot continue to develop medicines under these circumstances. The postdoctoral researcher who identified the alterations in DNA-modifying genes in leukemia in my lab is debating whether to continue his research. His bench mate, a talented chemist I hired out of graduate school, is teaching biology at a local night school to supplement her income. Next year, when she runs out of grant funds, she is thinking of returning to Florida to work in a tattoo parlor.

The Japanese people speak of a “Lost Decade”—the Ushinawareta Junen—a period between 1990 and 2000 when banks collapsed, the economy stagnated, and culture lost its effervescence. Emblematic of this decade was a “lost generation” of Japanese men and women who were unable to contribute to critical problems affecting their generation. My fear is that we too will face a lost generation—of molecules.

Cancer, meanwhile, marches onward. The statistics are stark: in the United States, one in two men and one in three women will encounter it. One in four will die from it. It’s important—given the complexity of the problem—not to oversell the speed or effectiveness of cancer research. Creating new medicines for cancer is slow, painstaking, time-consuming work. But that’s precisely why we need federal support to keep this process intact. It is exactly why—when Congress chooses to ax the NCI budget—we should take that decision with utmost seriousness. If we don’t generate enough political support for cancer research, we will not bring to life the kinds of medicines we need to treat our own cancers in the future—including the kind that killed Jobs. The vast gene-decoding efforts of the last decade will remain abstract, academic exercises—biology without medicine.

On my way back from the airport two days after his death, the signs and the apples were still outside the Apple store. I felt as if I should have added my own: “I’m sorry, Steve. I wish we had done better.”