The promise of young-blood injections to keep us sharp has captivated Silicon Valley, but the studies have all been with mice blood — until now. Researchers at Stanford University found that injecting mice with a protein found in human umbilical cord blood can improve their memory, though not everyone is as positive about the results.

Two years ago, the Stanford University lab of Tony Wyss-Coray, senior author of today’s study, showed that injecting young blood into old mice helped them perform better on memory tests. The team has moved on to injecting human blood of various ages, and the results are published in the journal Nature. After finding that old mice injected with umbilical cord blood performed better on tests, the scientists isolated a protein called TIMP2 in the blood that they thought was responsible for these effects — and then injected the protein by itself into more old mice and found the same positive results. It’s unclear how much this result would hold up in humans, but the team hopes this will be helpful for both academics and people working in the pharmaceutical industry, says study co-author Joseph Castellano, who researches neurology at Stanford.

But before you rush to the maternity ward of the local hospital, keep in mind that there are limits to the paper, says Irina Conboy, a bioengineer at UC Berkeley who has also studied young blood rejuvenation and published a less hopeful paper in Nature Communications last November. (Her paper showed that old blood makes mice old more quickly than young blood can rejuvenate.) The researchers suggest that our memory degenerates because we lose the TIMP2 protein as we age, and that’s why TIMP2-rich young blood could help us. But Conboy points out that previous papers show that there are high levels of TIMP2 in the cerebrospinal fluid of human patients with Alzheimer’s, a disease where your memory gets worse. “I am sure there is not just one silver bullet, and attracting attention to this idea that you can use human plasma injections is, I think, counterproductive,” she adds. “Furthermore, when you’re supported by a company that makes fresh-frozen human plasma, there is a very high temptation to say this is a miracle drug.” (The Stanford group is affiliated with Wyss-Coray’s company, Alkahest, which is studying the effects of young human plasma in Alzheimer’s patients.)

In the first step of today’s study, the researchers injected human plasma of different ages (umbilical cord, 19 to 24, or 61 to 82) into mice whose immune systems were destroyed so their bodies wouldn’t reject it. Mice who got the cord plasma did better on learning and memory tests, like escaping a maze. Mice who got the young adult plasma improved a little, and plasma from old people didn’t do much.

Next, the scientists used a device called a protein micro-array to analyze the amounts of different proteins in the blood. “We wanted to see what proteins were changing at various stages during human plasma aging,” says Castellano. “We wanted to see, what’s elevated in cord plasma that’s also changing in mouse plasma aging.” Using the arrays, they decided that the key protein was TIMP2, which regulates the activity of other proteins. There’s a lot of it in cord blood but not so much in older blood.

Next, they injected a lot of TIMP2 by itself into these old mice with a functioning immune system. The injections caused similar effects, according to Castellano, with these mice performing better on memory tests

But Conboy is more skeptical. First, the mice with destroyed immune systems are so sick to begin with that the good results might not be very representative. “When you say that something has been ‘rejuvenated,’ it has to be old to begin with, and more physiologically similar to people,” she says. “But the animals that they used are really very, very sick.” The immunosuppressed mice were about 12 to 14 months old, which is roughly analogous to the 40s or 50s in humans. It’s possible the memory “rejuvenation” was happening only because the mice were already sick, according to Conboy. She adds that at that age the results aren’t that impressive, since 50-year-olds don’t suffer severe memory loss and exercise and diet are usually enough to keep the brain healthy.

Plus, it’s not entirely certain that the results really suggest cord plasma could cause the same good results on normal old mice instead of ones with destroyed immune systems. The distinction between the enriched TIMP2 protein and plasma itself is key, given the number of companies that are focusing specifically on young plasma.

In the second part of the experiment, the researchers injected a lot of the purified TIMP2 protein into normal old mice. They labeled the TIMP2 protein and found that it does make its way to the brain. The problem is, they didn’t label the proteins in cord blood to see if TIMP2 was the key protein, according to Conboy. It’s a big protein, so it’s possible that TIMP2 is present in the cord blood but there’s not enough of it in plasma to cause big effects on normal old mice. Or, maybe TIMP2 from cord blood doesn’t make it to the brain, or maybe a different protein is the right one.

Castellano notes that it’s easy to label a single protein but much more challenging to label all of them in the blood and thus it’s impractical. In addition, when the researchers gave old mice plasma that didn’t have TIMP2, they didn’t see the same benefits, suggesting that TIMP2 really could be the protein creating these positive effects. Still, he emphasized that this is a pre-clinical trial and there’s still a lot to study. Next, the team wants to study the effects of TIMP2 in the brain more thoroughly, and also see how the protein affects mouse models of Alzheimer’s disease.