The One True Path?

Several groups of scientists are finding clues that suggest many major illnesses result from disruptions to one complex molecular cascade—insulin signaling.

ndocrinologist Kevin Niswender and neuroscientist Aurelio Galli hadn’t really kept in contact since they parted ways after beginning their respective careers at Vanderbilt University in the 1990s. But about 10 years ago, Niswender, who went to medical school at Vanderbilt, and Galli, who did a postdoc there, both landed faculty positions back at the Nashville, Tennessee, university. They rekindled their friendship and often discussed their research during convivial family dinners.

Niswender, who studies diabetes and metabolism, and Galli, who specializes in the neurobiology of addiction, had never collaborated scientifically. They can’t remember the exact moment they decided to do so, but gradually they realized that some of their research interests overlapped. The pair discussed...

Now, a decade later, Niswender and Galli are elucidating a molecular link between mental illness and problems with how the body processes sugars. That link is part of the complex series of events that make up the insulin-signaling pathway, a crucial mechanism by which the pancreatic hormone insulin directs the transport and storage of glucose in virtually every cell type in the body. This is only one of a recent rash of discoveries about how insulin is also intricately involved in many disease processes, including the growth of cancer cells and defects in bone mass regulation.

“The idea that insulin had effects independent of glucose uptake took a long time to be understood and recognized,” says Daniel Porte at the San Diego Health Care System, who saw the first hints that insulin signaling and brain function were somehow related in the early 1960s when he discovered that a class of brain hormones called catecholamines was controlling insulin secretion. “When we came up with that idea, that was considered pure heresy, because essentially everyone ‘knew’ that the only thing that regulated insulin was glucose.” Investigating further, Porte discovered that insulin was involved in how the brain regulates body weight and food intake.

Today, proposing a link between insulin signaling and disease processes previously thought to be unconnected to the pathway is less heretical. “I’m not surprised” that insulin may participate in a surprisingly wide range of diseases, says Johns Hopkins University biologist Thomas Clemens, who studies the link between insulin signaling and bone mass. “I think insulin has a broad role to play and I don’t think we’ve figured it out yet.”

Insulin, the hormone best known for its role in diabetes, is the body’s energy regulator. When functioning normally, pancreatic beta cells pump out insulin in response to increases in blood sugar (glucose) after a meal. Insulin instructs the body's cells to send glucose transporters to the cell membrane to absorb the sugar for the cell’s energy needs and convert the excess into energy-storage molecules, such as glycogen in the liver.

When insulin binds to insulin receptors in cell membranes, those receptors activate a number of insulin receptor substrate (IRS) proteins by phosphorylating them. Through these IRSs, insulin has an effect on many downstream actions, including regulation of glucose levels, lipid levels, and protein synthesis. Insulin receptors are expressed in every tissue in the body except mature red blood cells.

When insulin signaling gets disrupted, either because the hormone isn't being secreted (as in Type 1 diabetes) or because the cells don't respond normally to insulin (as in Type 2 diabetes), cells don't get the signal that the body has eaten, and therefore don't properly process sugars in the blood. Because diabetics can’t put the sugar they've ingested into storage, they suffer from sharp spikes in blood sugar levels after eating and intense low blood sugar when they haven't eaten.

In normal cells, a protein kinase called Akt is turned on downstream of the insulin receptor. Once activated, it phosphorylates four different proteins, each of which has downstream actions inside the cell—including making glycogen, lipids, and protein—and helping to push another protein, the glucose transporter, to the surface of the cell. Once there, glucose transporters shuttle glucose into the cell for processing. When Akt signaling is defective, those transporters remain in vesicles inside the cells, which cannot absorb any glucose. Akt malfunctions have also been linked to Type 2 diabetes and insulin insensitivity.1

But Akt also appears to be a key player in schizophrenia. Some schizophrenic patients exhibit impairments in Akt function,2 and many drugs used to treat mood disorders—such as lithium, and some antidepressants and antipsychotics—activate Akt by stimulating its phosphorylation.

Galli and other neurobiology researchers studying Akt noticed that in schizophrenic patients these disruptions decreased levels of the neurotransmitter dopamine in the brain’s prefrontal cortex. Dopamine deficiency in the prefrontal cortex is often a sign of mood disorders, including schizophrenia. Discovering how insulin, and specifically Akt, affects the brain’s dopamine levels has been Galli’s goal for more than a decade.

For his obesity studies, Niswender was breeding mice in which the function of Akt in neurons was blocked due to defects in its upstream pathway. The mice’s neurons lacked a protein called rictor (rapamycin-insensitive companion of mTOR), which forms a complex with mTOR (mammalian target of rapamycin). This complex, known as mTORC2 (mTOR complex 2), is activated by signals from the insulin receptor and the cell’s energy molecules, such as ATP. mTORC2, in turn, activates Akt and all of its downstream pathways (see Figure 1). Without rictor, the mice’s Akt pathway—and thus insulin signaling—was blocked in their neurons.

Figure 1). Normal neuron
In normal neurons, insulin binds its receptor and signals the initiation of several cascades, one of which activates a protein called Akt with help from a complex comprising the proteins mTOR and rictor (top of the cell). Activated Akt triggers additional molecular events, such as protein and lipid synthesis, cell growth, and glucose transport.

Here’s where Galli and Niswender put their dinner discussions to work in the lab. “We had a unique opportunity, now, using our techniques in his mouse, to explore more deeply what impairment in Akt and [what] insulin resistance in [the] brain really meant,” says Galli.

Looking closer at the Akt-deficient mice, Galli and Niswender noticed that just as in schizophrenic patients, the mice had lower brain levels of dopamine. But they also noted that the neurons in the mice’s prefrontal cortex had higher levels of cell membrane proteins called norepinephrine transporters, which bring dopamine and norepinephrine into cells. These overexpressed norepinephrine transporters pulled dopamine out of the synapse and back into the neurons, converting it to norepinephrine, thereby disrupting dopamine’s normal function as a neurotransmitter (see Figure 2).3 “By impairing Akt and increasing the number of [norepinephrine] transporters in the plasma membrane, you are creating a sort of vacuum for dopamine in the prefrontal cortex,” says Galli.

Figure 2). Akt-deficient neuron
In the prefrontal cortical neurons of Akt-deficient mice the disruption in Akt signaling increases the transcription of norepinephrine transporters. On the cell’s surface, these transporters vacuum norepinephrine and dopamine from the synapse into the cell (bottom of the cell), interfering with neuronal functioning and potentially causing some of the symptoms of schizophrenia.

The mutant mice were also more easily startled by sudden stimuli than normal mice, even when they received a warning signal prior to the stimulus. All of these changes are characteristic of the schizophrenic brain. Because the Akt pathway was only disturbed in the neurons of the mice’s prefrontal cortex and not in other tissue types, their systemic glucose, insulin levels, and sensitivity remained normal.

When Galli and Niswender treated mice with drugs that blocked the norepinephrine transporter, the mice returned to normal—their startle reactions to the same sudden stimuli were lessened by a warning signal, and dopamine levels in their brains returned to normal. Galli is currently trying to piece together the molecular mechanisms that connect mouse behavior with Akt and norepinephrine transporters in the brain.

Porte agrees that these studies show that Akt signaling is having an effect on the neuronal levels of norepinephrine transporters, but he thinks that the link between this and causation of schizophrenia is a big jump. “The studies are well done; I think the science is good,” says Porte. “When you get to ‘what does it mean to clinical disease?’ I think there you’ve got to be careful—a mouse is not a man.”

“I think this mouse and these experiments give us an opportunity to pick out one of the possible mechanisms [of how schizophrenia] is [destabilizing] dopamine signaling in the cortex,” says Galli. Understanding this pathway could help define unique approaches to treating mood disorders like schizophrenia, depression, bipolar disorder, and addiction, he says.

These findings may eventually help explain why depression, cognitive impairments and mood disorders are more common among diabetics, adds Niswender. “Understanding how insulin is working in the brain is a key piece of the puzzle in understanding how metabolism is balanced and how these [diabetic] comorbidities come about,” he notes.

“I think it’s a breakthrough, really. It’s incredibly impressive work,” says Zachary Freyberg, a psychiatry research fellow at Columbia University, who studies schizophrenia and the Akt signaling pathway. “[Galli] is basically combining several threads of evidence and weaving them together to create a pretty rich picture of the interplay between insulin-mediated signaling and some of the molecules…implicated in schizophrenia.”

There are also tantalizing molecular clues implicating insulin signaling in one of the world’s most studied diseases—cancer. Along with its many other metabolic functions, insulin also serves as a regulator of cell growth and proliferation, and if these functions are disrupted, there can be wide-reaching effects. “Insulin is a growth factor and cancer is a growing tissue, inappropriately growing,” says Porte. “[Cancer] pathology could involve an interaction with the insulin regulatory system.”

Because cancers arise from an amalgamation of different mutations, most tumors form due to changes in several different pathways. Because it’s crucial for normal growth and cell proliferation, the insulin signaling pathway, including the Akt pathway described in Galli’s schizophrenia work, is a source of tumor-promoting defects in many cancers, says cell biologist Brendan Manning of Harvard University, who studies how the Akt pathway is related to cancer.

Using transcriptional profiling to compare which genes get turned on and off when a cell becomes cancerous, a team led by Kevin Struhl, a Harvard Medical School geneticist, identified more 300 genes whose transcription is turned up or down when normal cells are transformed into cancerous cells. Among those, they found genes that play a role in lipid metabolism and metabolic diseases, including obesity, diabetes, and atherosclerosis. The genes come from a wide array of pathways, including insulin signaling and downstream lipid metabolism pathways.4

“There were various ideas out there [that these diseases are linked], but we are doing a very clear cancer-related project, and we come up with all of these obvious links,” says Struhl. “We had no idea what we were going to find. I mean, I was shocked when we found this.”

Struhl adds that there have been some “crude” epidemiological links between cancer and metabolism reported in the clinical literature, but the molecular mechanisms underlying these associations have yet to be fully determined. “I think that people hadn’t really thought about it that much and [my work] puts it on at least some form of molecular footing,” he says.

In conjunction with the experiment indicating that insulin signaling was playing some role in cancer progression, Struhl tested whether “drugs for one disease might work against another,” to see if common treatments for metabolic diseases, including diabetes, might be able to stop cells from becoming cancerous. Of all the drugs he tested on precancerous cells in vitro, the diabetes drug metformin had the biggest effect, slowing down the transformation of normal cells into malignant ones. This process normally takes a day, but when the cells were treated with metformin, they didn’t transform for over a week.

Struhl says that existing clinical data supports a link between metformin and cancer. In Type 2 diabetes patients, metformin interferes with malfunctioning insulin signaling pathways by activating a protein kinase that increases pancreatic insulin release and cellular uptake of glucose. Clinicians had noticed that diabetic patients often have higher cancer rates, but Type 2 diabetics taking metformin seemed to have lower cancer rates and improved cancer survival than diabetics taking other diabetes drugs. This suggests that metformin and insulin signaling possibly play roles in controlling and/or killing cancer (see Figure 3).

Figure 3). Cancer cell (left) Cancer stem cell (right)
Some of the molecular malfunctions in tumor cells occur within insulin signaling pathways. Tumor suppressor proteins, such as the tuberous sclerosis protein complex (TSC), are typically impaired, allowing for apoptosis suppression, uncontrolled cell growth, and increased protein synthesis—hallmarks of cancer. Researchers hypothesize that the diabetes drug metformin may be killing cancer stem cells (right) by stressing mitochondria, thus activating a protein called AMP-activated protein kinase (AMPK), which activates TSC. When this occurs, the function of a downstream protein complex that includes mTOR and raptor is inhibited, protein synthesis is scaled down, and apoptosis proceeds normally.

Struhl is investigating the hypothesis that metformin is acting to kill the tumor’s cancer stem cells, which give rise to new cancer cells and seem to resist the toxic effects of chemotherapeutic drugs. In recent experiments, his group found that when four different types of breast cancer cell cultures were treated with metformin, the drug specifically killed the stem cells in the culture.5

When Struhl combined metformin with the chemotherapy drug doxorubicin to treat the cultured cancer cells, the cocktail killed more cells than either drug alone, suggesting that they were working in complementary pathways. Furthermore, cancerous mice treated with both metformin and doxorubicin remained in remission longer than mice given doxorubicin alone. “If you treat with metformin and chemotherapy, the chemotherapy is nailing the traditional [cancer] cells and the metformin is killing the cancer stem cells,” says Struhl, adding that more work needs to be done to figure out the molecular mechanism through which metformin is acting to kill these cells.

I think insulin has a broad role to play and I don’t think we’ve figured it out yet. —Thomas Clemens

Metformin and other diabetes drugs are currently being tested in clinical trials, in combination with traditional chemotherapy, to determine if the cocktails treat cancers more effectively than chemotherapy alone. “I think it’s early,” Porte says, “early, but there seems to be something there. It’s safe to say that the full extent of the insulin signaling pathway, especially in its relation to cancer, is still up in the air.”

To witness how problematic disruptions to the insulin signaling pathway can be, look no further than diabetic patients themselves, says Gerard Karsenty, a developmental geneticist at Columbia University Medical Center. “If you look at patients who have Type 1 or Type 2 diabetes, it’s not only an increase or decrease in glucose blood levels. They have kidney diseases, eye diseases, reproduction defects, bone defects.”

Recently, clinicians have been documenting the extent to which one of these diseases, bone defects, is tied to disruptions in insulin signaling. This January, orthopedist Wojciech Pluskiewicz and his colleagues at the University of Silesia in Poland showed that adolescents with Type 1 diabetes have significantly weaker bones than their nondiabetic peers.6 A similar study of obese prediabetic adolescents found that they were more at risk for poor skeletal development.7 Type 1 diabetics are also more likely to experience early onset of degenerative bone disorders, such as osteopenia or osteoporosis.8

“In the clinic I’ve seen it. A lot of my diabetics, both Type 1 and Type 2 diabetes, have terrible bones,” says Clifford Rosen, a bone and metabolism specialist and endocrinologist at Maine Medical Center’s Research Institute. “It must be that the bone needs insulin, but we didn’t know how.”

While clinical evidence seems to point to a link between insulin and bone, only recently have studies begun to uncover the shared molecular root. The first hint came in the form of a protein called osteocalcin, which is made by bone-building osteoblast cells and plays a role in regulating bone mass. But osteocalcin has another important role—when it is decarboxylated, it acts as a hormone and signals the pancreas to secrete insulin.

To understand more deeply how insulin and diabetes affect bone mass and quality, a group led by Clemens at Johns Hopkins Medical School created mice that didn’t express insulin receptors on their osteoblasts, but still expressed the insulin receptor elsewhere in their bodies, including the muscle, fat, liver, and pancreas—the key players in whole body energy metabolism. When they studied these mice as they grew they noticed that, as expected from clinical and anecdotal evidence, the mice had low bone mass.9

What they didn’t expect was that, at about 10 weeks old, the mice started getting fat and became insulin resistant—just like Type 2 diabetics—suggesting that insulin signaling in bone was more important to systemic energy metabolism than previously thought. Clemens and his team were able to treat these diabetic symptoms with infusions of the hormone version of osteocalcin, which signaled the pancreas to pump out insulin.

The team found that, normally, when this insulin latched on to the insulin receptors of osteoblasts, it turned on a series of reactions that act to increase the production of osteocalcin (see Figure 4). It appeared as though the bones’ osteoblasts were relying on the insulin signal to continue secreting adequate levels of hormonal osteocalcin, and this hormonal osteocalcin was necessary for the body to respond to changes in glucose level. Disruption of this feedback loop might adversely affect both bone mass and metabolic regulation at the same time, says Clemens.

Figure 4). Bone cells
In bone-building cells called osteoblasts, insulin signaling inhibits the transcription factor Fox01, which increases the expression of a protein called osteocalcin (blue globes). Inhibiting Fox01 also downregulates a cytokine called osteoprotegerin (OPG, green triangles), stimulating the production and activity of the bone-degrading osteoclasts. As osteoclasts resorb bone tissue, the pH drops, converting osteocalcin into its hormonal form (red globes). This hormonal osteocalcin signals the pancreas to secrete more insulin. Studying disruptions in this insulin signaling cascade may help researchers better understand degenerative bone disorders, such as the osteoporosis and osteopenia that typically accompany Type 1 diabetes.

A group lead by Karsenty is also working with these mice to tease out a fuller picture of how insulin signaling and osteocalcin act upon the bone cells that form and degrade bone—the osteoblasts and the osteoclasts. Karsenty and his collaborators have found that normal insulin signaling in the osteoblasts not only increases the production and secretion of osteocalcin, but it also encourages the osteoclasts to degrade and resorb bone tissue. This resorption reduces the pH in the bone, which facilitates the decarboxylation of osteocalcin into its hormonal, insulin-stimulating alter ego (see Figure 4).10

To provide evidence that this isn’t just a mouse phenomenon, Karsenty investigated the osteocalcin and insulin levels in patients suffering from a disease called osteopetrosis, in which a reduced number of osteoclasts leads to a lack of bone resorption and very dense bones. These patients had significantly decreased levels of active osteocalcin and low serum insulin levels, even though their osteoblastic insulin receptors were intact. This supports Karsenty’s hypothesis that bone resorption is also an important step in insulin’s effects on bone.

Both Karsenty and Clemens say they can envision the potential impact of the insulin-osteocalcin loop on human health, because when insulin signaling is disrupted in bone, mice become systemically insulin resistant. “I think now one has to include bone in the equation of insulin target tissues,” says Karsenty.

“Nobody knew insulin and its receptor were the important component” of this feedback loop between insulin and osteocalcin and how it affected bone health and insulin sensitivity, says Clemens. Previous hypotheses focused on other metabolic hormones, such as leptin.

Fully understanding the connection between insulin regulation and bone quality is going to be important when treating diabetics, as aggressive treatment of their bone ailments might help treat their metabolic malfunction in the long run, says Rosen. “It’s obviously a cutting-edge area. The general concept is really exciting, that the skeleton mediates some metabolic activity,” he says. “Trying to understand what it does is really a challenge.”

“The clinic was telling us that insulin was obviously an important hormone—that’s an understatement—but was having many, many functions, not only regulating blood glucose,” says Karsenty.

“Almost all tissues have insulin receptors, so presumably insulin is having an effect on all those tissues. Because the dramatic effects on carbohydrate metabolism are so big, I think it has overshadowed all these other things that we are discovering now,” Porte notes. “It plays a major role in all other tissues in growth and development, and that’s why diabetic patients get into so much trouble.”

Have a comment? E-mail us at

1. A. Krook et al., “Insulin-stimulated Akt kinase activity is reduced in skeletal muscle from NIDDM subjects,” Diabetes, 47:1281-86, 1998.
2. D.L. Thiselton et al., “AKT1 is associated with schizophrenia across multiple symptom dimensions in the Irish study of high density schizophrenia families,” Biol Psych, 63:449-57, 2008.
3. M.S. Siuta et al., “Dysregulation of the norepinephrine transporter sustains cortical hypodopaminergia and schizophrenia-like behaviors in neuronal rector null mice,” PLoS Biology, 8:e1000393, 2010.
4. H.A. Hirsch et al., “A transcriptional signature and common gene networks link cancer with lipid metabolism and diverse human diseases,” Cancer Cell, 17:348-61, 2010.
5. H.A. Hirsch et al., “Metformin selectively targets cancer stem cells, and acts together with chemotherapy to block tumor growth and prolong remission,” Cancer Res, 69:7507-11, 2009.
6. A.P. Chobot et al., “Bone status in adolescents with type 1 diabetes,” Diabetologia, 53:1754-60, 2010.
7. N. Pollock et al., “Lower bone mass in prepubertal overweight children with pre-diabetes,” J Bone Miner Res, online ahead of print, DOI:10.002/jbmr.184, 2010.
8. V.M.G. Duarte et al., “Osteopenia: a bone disorder associated with diabetes mellitus,” J Bone Miner Metab, 23:58-68, 2005.
9. K. Fulzele et al., “Insulin receptor signaling in osteoblasts regulates postnatal bone acquisition and body composition,” Cell, 142:309-19, 2010.
10. M. Ferron et al., “Insulin signaling in osteoblasts integrates bone remodeling and energy metabolism,” Cell, 142:296-308, 2010.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?