A new study shows that these “regenerative medicine” clinics might be promising more than they can deliver.

Here’s the pitch. You’ve heard of stem cells, right?  They are these amazing cells that can change into any type of cell you may need based on where you stick them. They can become blood vessels, brain cells, cartilage-producing cells.  And here you are, with knee osteoarthritis that is just bone on bone. What I’m going to do is inject some stem cells into your knee, and they will take up residence there, and make more cartilage, and heal your pain.  Insurance won’t cover this.  I’ll charge you…say… $5000.

That pitch has led to a 35 Billion dollar industry – and a slew of stem-cell clinics popping up around the country making some really big promises. 

But if you look at the actual data – the actual trials of stem cell therapy for things like osteoarthritis – you will be very underwhelmed. Here’s a study showing the injections do no better than steroid injections (which are about 100 times cheaper) and which, frankly, don’t work that well in their own right.

In fact, you might come to the conclusion that these injections, in the best case, are – as my kids would say – pretty mid, and at the worst case just another variant of pseudoscientific nonsense. And this week, a new study has pushed me a bit more towards that pseudoscience interpretation.

I want to define my terms a bit here.  The study I’m reporting on this week is not about using stem cells in the context of bone marrow transplant or something like that.  This is really about those outpatient stem cell procedures – done in the office – to relieve various aches and pains.

Patients typically come to a “regenerative medicine” office and, after a consultation, have what amounts to a small bit of liposuction to harvest fat from the abdomen or, less commonly, a bone marrow aspirate. Both sources of tissue are thought to be rich in stem cells, and are used interchangeably. The aspirated tissue is not processed – the FDA has not approved the use of cultured cells in these procedures – so basically the fat is sucked out of your belly and injected into your knee.  That’s the procedure.

And so right away you have to ask yourself – is what I got stem cells?  I mean, clearly, you got more than stem cells, you got fat cells and probably some blood and plasma, and various proteins and other substances.  So maybe the right question is – ok how many stem cells did I just get?

And that’s where this study, appearing in Science Advances, takes off.

Researchers enrolled 21 participants and performed the same procedure happening at 100s of clinics around the world – they aspirated some bone marrow and some abdominal fat – two supposedly interchangeable sources of stem cells.  But instead of injecting them into knees, they injected them into some of the most sophisticated analytic machinery a modern molecular biology lab has to offer.

Let me start with the first, biggest question. If you get an injection of fat from your belly into your knee – how much of that will be stem cells?

Turns out that there are some standard ways to define whether a given cell is a stem cell or not. These are based on markers that appear on the surface of the cell. The unique combination of the various markers tells us what kind of cell it is in great detail.  One combination represents a T-helper cell, one a fibrocyte. What we’re looking for is mesenchymal stem cells, which have this particular surface marker combination:

In sequential steps, filtering the data, the researchers probed the bone marrow slurry for these stem cells. And they found 3.5.

Not 3.5 million. Not 3.5 thousand. 3.5 cells. Per sample.

In one of the samples, there were no stem cells at all.  In the sample with the highest concentration, 0.017% of cells identified were stem cells. The number was so low that the researchers couldn’t even rule out the possibility that all these cells were just false positives.

The fat sample did a bit better, with a median of 1.7% of identified cells appearing to be stem cells. But when those cells were examined, they found them to be of various different sizes and textures – not a uniform field of cells as they should be.

This also suggests a high false-positive rate. Maybe these preparations have basically no stem cells in them?  Or maybe the definition of stem cell is wrong. The authors propose the latter, and decide to try another way to find the elusive stem cells.

Enter tissue transcriptomics. Here, the RNA from the cells harvested from fat or bone marrow is analyzed. So instead of asking what the surface markers of the cells are, we are asking a more fundamental question – what are these cells doing? What stuff are they making? 

The answer, of course, is a lot. But by measuring all that stuff, we can map how similar one cell is behaving to another in a plot like this.  Each dot here is a cell – black are cells from bone marrow and pink from fat.

Note, first of all, that pink and black are not mixed together here. What that means is that the cells from bone marrow and from fat are very different – not interchangeable as has been suggested in the past.

And look down there… in that little red box. Here are the only cells that appear to be stem cells.  A tiny fraction – and all black, by the way – only present in bone marrow, not fat. But lost in a sea of other cells doing other things.

Another experiment enriched the slurry with cultured stem cells – these guys are the real deal – the gold standard – and no you can’t get them – cultured stem cells are not FDA-approved for injection for knee pain or anything else at this time. You can see there’s a bunch of them, in blue, and that they are kind of close to the smattering of bone marrow stem cells, but not overlapping.

Looking at this, I would argue that, if these injections work (and even that is a big if), they do not work because of stem cells.  There is a lot of other stuff in there.

So far – we’ve only talked about cells. That’s not all you get when you stick a needle into a fat pad and suck some out. You get a lot of protein as well.  And the researchers characterized all the proteins in the fat and bone marrow aspirates.

Of note – only ten of the proteins detected had some significant immunomodulatory function that might aid in healing a bad joint. And only four of the ten were detected in both fat and bone marrow aspirate.

This is an incredibly detailed paper – which comes with a fully interactive website too – so I encourage anyone interested to take a deeper look.  But I think what I want to highlight here is that, basically, we have no idea how these “stem cell” injections work, if they work at all. And the reason for that is that when you inject a whole bunch of stuff into a joint – cells, proteins, fluids, cytokines, and so on – it becomes impossible to tell which substance is biologically active. Maybe the reason the trial data on these therapies is so all over the map is because the stuff your injecting can be wildly different based on where you got it from. 

Yeah, it’s not clear to me whether or how these injections work.  What is clear to me is that the marketing of “stem cell therapy” and “regenerative medicine” absolutely works. It tells a great, simple story – harvesting your own body’s cells to heal yourself. No medications – no surgery. Just a bit of fat, a bit of hope, and $5000 that your insurance won’t cover.

The idea, except the cost, is a nice one. But I think if we are being honest and really believe that stem cells are the magic bullet here, then we need to either isolate the stem cells prior to injection, or, ideally, culture them to have enough of them to actually make a difference. Neither cell isolation or culture for this purpose is FDA-approved now. To get approval, someone will need to show that their preparation actually works. If the marketing is right, that shouldn’t be hard at all. Right?

A version of this commentary first appeared on Medscape.com.