Translational Stroke Research

, 2:266

Intracranial Delivery of Stem Cells

Authors

    • Institute of Neuroscience and PsychologyUniversity of Glasgow
    • Institute of Neurological Sciences, Southern General Hospital
  • John Sinden
    • ReNeuron
  • Erik Miljan
    • ReNeuron
  • Laurence Dunn
    • Institute of Neurological Sciences, Southern General Hospital
Cell-based Therapies for Stroke

DOI: 10.1007/s12975-011-0095-z

Cite this article as:
Muir, K.W., Sinden, J., Miljan, E. et al. Transl. Stroke Res. (2011) 2: 266. doi:10.1007/s12975-011-0095-z

Abstract

The method of delivery of stem cells is a major factor to consider in the design of clinical trials of cell therapy. Different methods of delivery will be associated with different risks to the patient, and may also be associated with different potential for benefit. Current approaches are partly informed by the routes selected for study in animal models of focal ischaemia and CNS transplantation, but there has been little work comparing the efficacy of different routes of administration. Direct intraparenchymal delivery of cells has been employed in several preliminary clinical trials, and data on the safety of this approach are reviewed.

Keywords

Stem cellStrokeIntracranial injectionClinical trialSafety

Background

Clinical trials of cell therapies face a large number of design issues, for most of which there are presently only limited data—and frequently none at all—to inform investigators. These include patient factors (age range, interval since stroke, stroke type), factors related to measurement of efficacy (nature and severity of neurological deficit, adjunctive use of brain imaging or clinical assessment tools), the validity of different control groups (historical, concurrent ineligible, sham) and issues that primarily relate to mode of delivery [1].

The small number of clinical studies that have been undertaken to date have employed either direct intracranial delivery of cells via stereotaxic neurosurgery (including teratocarcinoma-derived NT2 cells [2, 3] and the current trial of the CTX0E03 neural stem cell line—the PISCES trial—http://www.clinicaltrials.gov/ct2/show/NCT01151124?term=PISCES&rank=2) or intravascular delivery of bone marrow-derived cells (cells with surface markers of mesenchymal stem cells selected and expanded ex vivo given intravenously [4] or current trials of CD34+ cells delivered by the intra-arterial route). The former studies using direct intracranial delivery have been safety evaluations in patients late after stroke (minimum of 6 months after the incident event), while the latter have delivered cells in the early subacute phase within weeks of the event. These trial designs do not follow logically from the cell types involved, since allogeneic cell grafts should permit “off the shelf” use in patients without delay, while the need for ex vivo expansion of autologous cells derived from bone marrow necessarily entails a delay of several weeks before administration is possible. Rather, the trial designs have been dictated by the mode of cell delivery that was selected, for example, intracranial delivery often requires general anaesthesia, stability of brain anatomy to minimise risk from stereotaxic delivery and must be planned around the temporary discontinuation of antithrombotic drug treatments, all of which lead to a preference for inclusion later after stroke. Intravascular administration has the hypothetical disadvantage that cells may not be able to cross an intact blood–brain barrier in adequate numbers for therapeutic effect, therefore favouring earlier time points after stroke when this is likely to be damaged. In addition, there is no control over the anatomical target to which the cells are delivered.

Animal Model Evidence

Animal models of focal ischaemia and cell transplantation underpin the clinical trial approaches that have been followed. Neural stem cells have been delivered by a mixture of direct stereotaxic injection, predominantly into the putamen or striatum ipsilateral to the infarct, although ipsilateral cerebral cortex, intracisternal, intravenous (IV) or intra-arterial (IA) routes have also been studied [5]. The interval between induction of focal ischaemia and cell implantation has been up to 4 weeks. Mesenchymal stem cells or other bone marrow-derived cells have been delivered by intraparenchymal injection, IV or IA routes [610].

No direct comparisons of efficacy determined by behavioural testing have been done with the same cell type delivered by different means, and interpretation of studies where cells have been delivered by different routes as indirect comparisons are of questionable validity since other experimental conditions have been varied (e.g. ischaemia model, interval from ischaemia to implantation, behavioural test battery). In one possible indirect comparison, rats implanted with human teratocarcinoma-derived hNT cells in brain cortex adjacent to the site of infarction failed to exhibit consistent improvements in neurological function despite survival and neuronal differentiation being evident histologically [11], whereas the same cell line had earlier been associated with significant behavioural recovery when implanted into ipsilateral striatum in a rodent focal ischaemia model [12].

Clinical Studies

The history of implantation of cells into the brain goes back to the early 1980s. Intracerebral stereotaxic cell delivery into the brain has been employed in stroke, and also for other neurological diseases, notably Parkinson's disease (PD) and Huntington's disease (HD). Older studies, including adrenal medulla implants and foetal implant studies, were reviewed by Rehncrona [13]. On behalf of the American Academy of Neurology, Hallett and Litvan reported the evaluation of clinical studies in patients with PD in which a range of experimental functional surgical procedures were employed [14]. Their summary of the evaluation of implantation of human foetal ventral mesencephalon tissue and cells concluded that foetal tissue transplants were well-tolerated and that morbidity and mortality were low. Occasional morbidity included transient confusion after surgery and small haemorrhages near the implantation or needle-track sites. However, in cases where long-term immunosuppression was undertaken, opportunistic pneumonia did occur [15].

Commercial sponsors in the late 1990s and early 2000s supported trials in the USA, notably Diacrin Inc, using porcine-derived primary foetal tissue with early phase trials in PD [16], HD [17] and disabled stroke patients [18]. Layton Biosciences Inc. used a human teratocarcinoma cell line to derive differentiated neuronal-like cells (hNT neurons) and undertook phase I/II and phase II trials in stable stroke patients [2, 3]. In addition, the NIH funded two larger-scale phase II controlled trials in the USA using implants of human foetal ventral mesencephalon tissue strands and pieces [19, 20].

In order to examine general procedural adverse events, we reviewed published studies from 1990 to May 2011 that involved human and porcine primary foetal brain cells and tissues and manufactured cell lines and that reported sufficient detail of procedures and complications. English language publications were identified using freely available search engines (PubMed and Google Scholar) with the terms Cell and Therapy and Clinical Trial and Stroke/Parkinson's disease (PD)/Huntington's disease (HD). We excluded studies of autologous tissues and cells (e.g. adrenal medulla and bone marrow). Where possible, reported serious adverse events (SAEs, defined to be life-threatening, necessitating hospitalisation, causing persistent or significant disability/incapacity or death within 1 year post-operatively) were separated into those due to the surgical procedure, the implanted cells or other aspects of the trial procedure, for example immunosuppressive drug regimes. Details of clinical reports of intracranial delivery of cells for PD and HD are given in Table 1, and for stroke in Table 2.
Table 1

Clinical studies of intracranial delivery of cells for Parkinson's and Huntington's diseases

Reference

Description

Patients (n)

Tracts per side

Estimated cells/tract

Volume per tract

Immunosuppression

Surgery

Graft

Other

Parkinson's disease

Wenning 1997 [32]

Unilateral, cell suspension

10

3–5 (PU)

0.8–3.5 × 106

NR

C,A,P continuous

No

No

No

Hagell 1999 [33]

Second contralateral on patients from (Wenning et al. 1997)a

4

5 (PU)

1.2–2.1 × 106

20 μl

C,A,P continuous

No

No

No

Brundin 2000 [34]

Bilateral, cell suspension with Lazaroid treatment

5

5 (PU), 2(CN)

0.66–1.08 × 106

NR

C,A,P (12–24 m)

No

Nob

Yesc

Peschanski 1994 [35]

Unilateral, cell suspension

2

3–4

1.5–2.25 × 106

24 μL

C,A,P (7 m)

No

Nod

No

Defer 1996 [37]

Unilateral, cell

5

4 (3 PU), 1

1.4–1.9 × 106

24 μl (8 × 3 μL)

C (7 m), A,P

Noe

Yesf

No

Jacques 1999 [38]

Summary of multiple trials

60

4

NR

Volume of graft 20–24 mm3

NR

Yesg

Yesh

No

Schumacher 2000 [16]

Porcine cell xenograft unilateral

12

3

4.0 × 106

80 μL (one patient 200 μL in two sites)

C (6 patients) anti-MHCII (6 patients)

No

No

No

Freed 2001 [19]

Double-blind randomised trial–bilateral grafts

19

2i

1.5 × 106

20 μL (constant delivery upon withdraw al)

None

Yesj

Yesk

No

Mendez 2000 [36]

Bilateral, putamen cell suspension, stored in GDNF

2

4

1 × 106

10 μL (4 × 2.5 μL deposits)

C (6 m)

Nol

No

No

Olanow 2003 [20]

Double-blind placebo-controlled trial

11/12

8

0.19–0.75 × 106

NR

C (6 m)

No

Yesm

No

Henderson 1991 [39]

Caudate, unilateral large grafts

12

1

1.5 × 106

0.5–2.0 mL

None

No

No

No

Huntington's Disease

Sramka 1992 [40]

Foetal striatal suspension

4

2–3

NR

NR

C

No

No

No

Madrazo 1995 [41]

Foetal striatal tissue pieces

2

Ventricular wall

NA

NA

C, P

No

No

No

Kopyov 1998 [42]

Bilateral foetal striatal suspension graft

3

1 (CN) +4 (PU)

NR

NR

NR

No

No

No

Fink 2000 [17]

Porcine xenografts

12

2 (CN) + 4 (PU)

2–4 × 106

NR

C

No

No

No

Hauser 2002 [43]

Bilateral foetal striatal suspension

7

PU

NR

NR

C (6 m)

Yesn

No

No

Bachoud-Levi 2000 [45]

Unilateral foetal striatal suspension

5

2 (CN) + 3 (PU)

NR

40 μL (8 × 5 μL deposits)

C,A,P (1 year)

No

Noo

Yes

Rosser 2002 [30]

NEST-UK Trial (unilateral foetal striatal suspension)

4

2 (CN) + 4 (PU)

NR

12 μl (5 × 2.4 μL deposits)

C,A,P (6 m)

No

No

Yes

C cyclosporine A, A azathioprine, P prednisolone, PU putamen (including anterior, medial and posterior), CN caudate nucleus, NR not reported, NA not applicable

aResults of a second implant, 10–56 months after the first with four to eight donor tissues

bCognitive events reported such as mild confusion and depression in two patients that resolved with time or treatment

cPatients also received tirilazad mesylate treatment to improve graft survival which gave rise to peripheral thrombophlebitis in all patients

dTransient postoperative period of heavy dyskinesias, alleviated with l-Dopa dose adjustment

eTemporary frontal syndrome developed in three patients and one patient developed transient confusion—both resolved within days after surgery

fAll patients experienced some degree of bilateral dyskinesias which improved in months following surgery. Three patients developed dyskinesias contralateral to the graft 2 years post-surgery

gNine cases of complications out of 60 patients (15%), including haematoma (5%), frontal lobe syndromes (3.5%), frontal lobe syndrome and hematoma (1.5%), cyst formation (1.5%) and seizures (1.5%)

hOne patient exhibited severe dyskinesia which could not be resolved by reducing l-Dopa levels

iSurgical details differ in this study from others in that needle trajectories were made from a frontal approach

jNine serious adverse effects reported, only one of which (subdural haematoma) was a possible consequence of the surgery

kSevere dystonia and dyskinesia at 12 months in 5 of the 18 younger patients that did not respond to reductions in l-Dopa. The remaining six patients were advised against the treatment based on the severe uncontrollable dyskinesias in younger patients and the lack of apparent efficacy in older patients and the trial was discontinued

lSurgical details reported (Mendez et al. 2005 [44]). Patient 2 left hemisphere graft could not be completed due to bleeding through cannula—no neurological deficit related to the complication during surgery

mFifty-six percent of transplanted patients developed dyskinesia that persisted after overnight withdrawal of dopaminergic medication (off-medication dyskinesia)

nThree subjects developed subdural haemorrhages and two required surgical drainage. One subject died 18 months after surgery from probable cardiac arrhythmia secondary to severe atherosclerotic cardiac disease

oDirect physiological consequences of the of the cell graft itself were rare and not worrisome, although mood alterations in patients and relatives did require medical treatment

Table 2

Clinical trials of cell therapy delivered by intracranial injection in stroke

Reference

Description

Patients (n)

Tracts per side

Estimated cells/tract

Volume per tract

Immunosuppression

Surgery

Graft

Other

Savitz 2005 [18]

Porcine LGE xenotransplant

5

3–5

10 × 106

100 μl

No

No

Yes

No

1

80 × 106

800 μl

Kondziolka 2000 [3]

LBS neuron

8

1

2 × 106

60 μl

Cyclosporine (6 m)

Yes

No

No

4

3

2 × 106

 

Konziolka 2005 [2]

LBS neuron

7

5

1 × 106

50 μl

Cyclosporine (6 m)

   

7

5

2 × 106

C cyclosporine A, A azathioprine, P prednisolone, PU putamen (including anterior, medial and posterior), CN caudate nucleus, NR not reported, NA not applicable

In summary, there are data relating to 238 patient surgeries involving cell grafts into the brain. Most were allogeneic primary foetal brain tissue or cell dissociates of variable quality, but two trials used manufactured cell lines, and three trials used porcine xenograft cells. The number of needle trajectories made per side of the brain was between one and eight, and bilateral surgery, where appropriate, was carried out in the same session in the majority of cases. The cell number injected is not available in most cases where foetal-derived brain tissue was used. When reported, the number of cells injected per tract was ∼0.5 to 10 million cells in volumes of 10–200 μl per tract, with injection bolus size from 3–100 μl. In one case, 80 million cells were delivered in 800 μl in a single tract. In terms of surgical risk, approximately 12 incidences of SAEs directly attributed to the surgery were reported. This frequency is consistent with analysis of safety data reported for stereotaxic brain procedures undertaken for both functional (pallidotomy, electrode placement) and morphological (biopsies, evacuations) stereotaxic neurosurgery. Reported in the literature are an overall risk of bleeding of 1.7% for any type of stereotaxic procedure, resulting in a mortality of 0.3% and a morbidity of 1.4%. SAEs typically included subdural hematomas and seizures [21]. Following stereotaxic electrode placement, another study of 337 patients showed a symptomatic haemorrhage rate of 1.2%, with a rate of permanent neurological deficit of 0.7% [22], where haemorrhages were defined as symptomatic if they were detectable on CT or MR imaging and changes were observed on neurological examination. Previously published risks of symptomatic haemorrhage, ranging from rates of 0.6–2.0%, following stereotaxic surgery, were consistent with these reports [2326]. The number of needle passes into the basal ganglia may increase the incidence of deficits post surgery. The risk of intracranial bleeding, during or after stereotaxic surgery has been reported to be greater in patients with additional risk factors, (e.g. vascular fragility, various coagulopathies or unstable blood pressure) [21] or hypertension in the intraoperative or postoperative periods (systolic blood pressure 160 mmHg or higher) [22].

SAEs that may be related to cell implantation were reported from 2001 in the PD studies where implants increased dopaminergic function, including dyskinesias and hallucinations [27]. In particular, the placebo-controlled trials of human ventral mesencephalon tissue [19, 20] produced unacceptable levels of dyskinesias, with 56% of the treated patient group experiencing disabling off-medication dyskinesias in one trial [20]. It would appear that there is a general moratorium on clinical trials of foetal tissue transplants in PD patients, awaiting developments, currently being tested preclinically, that would limit dyskinesia side effects [28].

Clinical trials of human foetal tissue transplants in HD are, on the other hand, continuing with no apparent evidence of frequent safety concerns (see Freeman et al. [29]. for a detailed review)

Side effects and complications considered to be related to immunosuppressive treatments were reported in two HD trials, including derangement of renal function, and hirsutism related to cyclosporine [30], that caused the investigators to question the use of immunosuppressive drugs in this type of cell therapy.

Stereotaxic injection of cells has used specially designed injection devices to deliver small but controlled volumes. The cannula design employed for the clinical hNT neuron trials has been described in detail [31] and a similar device is being used in the current PISCES trial.

The details of cell delivery in three previously reported clinical studies in stroke patients are summarised in Table 2. Savitz and colleagues [18] reported no perioperative complications; however, one patient received 800-μl delivery of cells in less than 1 min via a second burr hole as the first did not provide an adequate trajectory for cell implantation. This trial was terminated after the inclusion of five patients due to adverse events in two subjects: patient 5 developed generalised and partial complex seizures, while hyperglycaemic and experienced a number of episodes of diabetic ketoacidosis over a 4-year period; patient 4 experienced increased weakness of the left arm 20 days after surgery which was deemed to be related to cortical vein thrombosis likely secondary to the surgical procedure. In the two studies by Kondziolka and colleagues [2, 3], doses of 2 to 10 million cells were delivered by multiple needle trajectories and multiple injection boluses of 10–20 μl each at points along each needle pass into the basal ganglia. There were two procedure-related events post surgery: one patient developed a single seizure 1 day after surgery, while another patient (taking dual antiplatelet therapy) developed an asymptomatic subdural haematoma noted on follow-up magnetic resonance imaging that was treated by burr hole drainage 1 month after surgery.

Conclusions

There is considerable experience with stereotaxic neurosurgical intracranial delivery of cell therapies, with procedural complication rates that are similar to those reported for other stereotaxic interventions, principally intracerebral haemorrhage in 1–2% of patients. Effects of grafted cells themselves are only evident in the development of dyskinesias in patients with PD, and other complications have occurred mainly in relation to adjunctive immunosuppressant therapy.

Stereotaxic intraparenchymal delivery offers some advantages in cell therapies for stroke, ensuring that large numbers of cells are adjacent to the site of ischaemic tissue damage, and avoiding any concern that cells may fail to cross the blood–brain barrier. This opens the possibility for later intervention following stroke, beyond the subacute phase when blood–brain barrier permeability is increased. Safety trials to date have delayed intervention in order to ensure patient stability and fitness for anaesthesia, but there is no intrinsic reason why stereotaxic procedures could not be undertaken earlier. Delivery by intraparenchymal injection may impose some anatomical restrictions since some sites are unlikely to be safely amenable to stereotaxic injection (e.g. brainstem) or may be inaccessible due to cerebromalacia.

Copyright information

© Springer Science+Business Media, LLC 2011