I n late 2015, cell therapy company Stem- Cells Inc. offered encouraging news for people with debilitating spinal cord injuries. In early results from a trial known as the Pathway Study, an injec- tion of neural stem cells, which can form
neurons and other cell types, seemed to be
reversing some effects of spinal injuries in
four of six patients. Yet 6 months later, the
company declared it was shutting down.
An expanded group of patients wasn’t regaining movement and strength as hoped.
StemCells Inc. merged with sur-
gical device company Microbot
Medical Inc. last fall, and hasn’t
published results from the study.
Now, some of the company’s
longtime academic collaborators
have come forward with disheartening animal data, and an admonition for future stem cell clinical
trials. A team at the University of
California (UC), Irvine, found that
a clinical stem cell line provided
by StemCells Inc. produced no
benefits in mice with upper spinal cord injuries—a discovery the
researchers say they shared with
the company before the Pathway
Study’s launch. A second mouse
study, published by another UC Irvine team in the same 14 February
issue of Stem Cell Reports, found
no benefit from a different neural
stem cell line the company was exploring to
treat Alzheimer’s disease, but that hadn’t advanced to clinical trials.
More and more companies are developing
stem cell therapies, but the UC Irvine teams
caution that their results show how large-
scale clinical manufacturing of cell lines—a
process often guarded as a trade secret—can
change stem cells in unpredictable ways.
They and others are concerned that regula-
tors don’t require sufficient animal testing
of the exact cell lines destined for patients.
The results “may give a warning for future at-
tempts by similar companies,” says bioethicist
Insoo Hyun at Case Western Reserve Uni-
versity in Cleveland, Ohio. “Hopefully these
papers will encourage careful efficacy studies
in animal models before going to humans.”
StemCells Inc. developed one of the first
in a slew of potential spinal cord therapies
involving various kinds of stem cells, which
have yielded a few promising—if anecdotal—
results from small clinical trials. Starting in
2002, the company worked with the labs of
UC Irvine neurobiologists Aileen Anderson
and Brian Cummings to test human neural
stem cells, derived from fetal tissue, in mice
with injuries to the midspine, known as the
thoracic region. Based on encouraging rodent
results, the firm then conducted a clinical
trial in Switzerland and Canada on 12 peo-
ple with that type of injury, and announced
that seven regained some sensation a year
But damage to the cervical spine—the
region at the back of the neck—is more
prevalent, and often more debilitating. In
the summer of 2013, as StemCells Inc. prepared for a trial targeting such injuries, the
UC Irvine group won a grant from the U.S.
National Institutes of Health to test the company’s cells in a mouse model of cervical
spinal damage. This time, they would look
not just at research-grade cells grown for lab
studies, but also at a line of clinical-grade
cells planned for use in people. These lines
are typically made in larger batches under
stricter manufacturing conditions.
The clinical-grade cells that arrived from
the company formed large, sticky clumps
with more debris and fewer viable cells than
the researchers were used to, Cummings
says. And they yielded disappointing results.
Compared with the research line, fewer of
the injected clinical cells survived after trans-
plant, and fewer appeared to differentiate
into oligodendrocytes, supportive cells that
can help insulate neurons and improve nerve
signaling. Neither cell type led to dramatic
improvements in the rodents’ motor skills;
the clinical cells even seemed to reduce
treated mice’s performance in tests of gait
The researchers say they showed these
preliminary results to StemCells Inc. in July
2014, and were surprised when the company
still launched its clinical trial. “There’s no
ethical way you could go forward with this
trial in people,” Cummings says.
He also worries that patients
couldn’t have provided proper in-
formed consent for the trial. “They
were getting cells that there was
no published data on,” he says.
In a letter published alongside
the papers, StemCells Inc. denies
that the animal results warranted
any change to the human trial.
The cervical injury mouse model
may not be predictive of effects
in humans, it suggests. The com-
pany also says it shared the UC
Irvine results with study clini-
cians and with the U.S. Food and
Drug Administration (FDA).
The letter’s corresponding au-
thor, a lawyer representing the
company, did not respond to
requests for further comment.
Hyun, who helped draft guidelines for developing cell therapies, says it’s
surprising that FDA apparently greenlighted
the clinical trial without evidence of efficacy
for the specific cell product headed to the
clinic. According to an FDA spokesperson,
“animal studies that use the clinical-grade
product are generally preferred.” But the
agency doesn’t require them, in part because
these human cell products can provoke immune reactions in mice, or require immune
suppression that complicates results.
No animal model can perfectly predict what a treatment will do to humans,
Anderson acknowledges, but she hopes her
team’s findings will point out an under-appreciated culprit in clinical trial failures.
Often, “what happens on the clinical side
doesn’t really match the question asked on
the basic side,” she says, “and that’s a crazy
way to go forward.” j
Neural stem cells, shown migrating out of a cultured sphere, show promise for
repairing spinal cord injuries, but are sensitive to variations in manufacturing.
By Kelly Servick
Failed spinal cord trial offers cautionary tale
Collaborators contend clinical stem cell lines lacked support from animal studies
NEWS | IN DEPTH