Method of long-term, recurrent, intracerebroventricular infusion of cellular therapy in mice

1. Introduction

Central nervous system (CNS) diseases, including brain tumors and neurodegenerative diseases, remain significant health challenges in part because the blood-brain barrier largely prevents therapeutic agents from reaching the CNS. Treatment delivery via direct access to the CNS provides a substantial advantage because it allows escalation of the therapeutic dose while limiting systemic toxicity. Therefore, clinicians regularly use intracerebroventricular (ICV) devices such as Ommaya reservoirs to deliver high-dose chemotherapy and antibiotics to patients who have tumors and CNS infections.¹ In clinical trials, these devices are now being evaluated for their capacity to deliver cellular therapies to the CNS of patients with brain malignancies and neurodegenerative diseases.^2,3

The development of cellular therapies for CNS malignancies and neurodegenerative diseases, especially those delivered with ICV devices, is limited by current methods of assessing these therapies preclinically. For example, patients with the neurodegenerative disease amyotrophic lateral sclerosis tolerated stem cell implantation well via the Ommaya reservoir, owing to the reservoir’s ease of repetitive injections⁴. In ongoing clinical trials of chimeric antigen receptor (CAR) T cells for glioblastoma (e.g., NCT02208362 and NCT04003649),² the clinical trial plan also includes treating patients with multiple ICV injections of the therapy; however, the prior preclinical investigation in glioblastoma-bearing mice was limited to a single ICV administrations of CAR T cells.⁵ The preclinical study showed that CAR T cells are effective against glioblastoma and that ICV treatment requires fewer cells than intravenous infusion to achieve a similar therapeutic effect⁵; but it did not provide reliable evidence for scheduling repeated ICV treatments in humans. Thus, preclinical strategies that enable repeated administrations of cellular therapies to the CNS, which would more accurately reflect treatment schedules in the clinic, are needed to accelerate such therapies’ transition from the bench to the bedside.

An improved method to deliver cellular therapy into the CNS may also help overcome many of the technical impediments limiting the success of cellular therapeutics for ameliorating pathology across pre-clinical models of brain diseases. For example, CAR myeloid cell therapy, like CAR T cell therapy, is effective against cancer but is difficult to deliver⁶. Specifically, the delivery of myeloid cells to the brain via peripheral infusion results in a significant entrapment of myeloid cells in the peripheral organs, and the number of cells that reach the brain diminishes after 1 week⁷. Thus, maintaining a therapeutic level of cells in the brain requires frequent infusions⁷. These inefficient approaches relying on repeated peripheral infusions are used due to the lack of a delivery system that can reliably deliver myeloid cells directly and durably to the brain. Methods of repeated intrathecal injections for the delivery of such cellular therapies in rodent models have been developed; however, they are used mainly for pathologies of the spinal cord, not the brain.^8,9 Another preclinical tool, the Alzet Brain Infusion Kit, offers direct access to the cerebral ventricles, and its osmotic pump provides a constant and slow release of the selected agent¹⁰. However, cellular therapies, unlike drugs, do not remain stable in the pump for long, and their viability and efficacy quickly decay.

Thus, the advancement of cellular therapy for CNS diseases requires an improved surgical protocol for the continuous or periodic delivery of cellular therapeutics to the brain in small rodent models. To our knowledge, no established surgical protocol meets this need. Accordingly, we developed and validated a minimally invasive surgical protocol that enables the long-term, repeated ICV delivery of cellular therapy in mice.

2. Materials and methods

2.1. Materials

Either 1) an Alzet Brain Infusion Kit 2 or 3 (DURECT, 0008663 and 0008851) or 2) an ICV device with a 2.5-mm intracranial needle (PlasticsOne, 3280PM/SPC) connected to a polyvinyl chloride (PVC) catheter or a silicone catheter (PlasticsOne, CT24SR) was installed in male C57BL/6J mice (Jackson Laboratory) for the demonstration and validation of the method, or male P301S mice (Jackson Laboratory) for the comparison of complication rate between methods. Four 0.5-mm-thick spacers were installed with Kit 2 so that the intracranial needle is 3.0 mm as in Kit 3. Superglue (Cyanoacrylate, Pacer Technology) was used to fix the ICV device to the skull. A Hamilton syringe with a 22G flat tip (Hamilton Company, 80285) was used to infuse cells or phosphate-buffered saline (PBS). The PVC catheter was sealed using a cauterizer (Symmetry Surgical, DEL1), and the silicone catheter was plugged with the end of a P10 pipette tip (Thermo Fisher). The surgical wound was closed using 5–0 nylon sutures or the Autoclip wound closure system (Becton Dickinson, 427638).

2.2. Generation of luciferase-expressing monocyte cell line.

Retroviral vector pMG-Lyt2-Luciferase was used to transduce the monocyte cell line PMJ2PC (ATCC, CRL-2457). The successfully transduced cells were enriched using mouse CD8α MicroBeads and an LS column (Miltenyi Biotec, 130–117-044 and 130–042-401). The stable expression of the transduced construct was assessed on an LSRII flow cytometer (BD Biosciences).

2.3. ICV device preparation

A 15.0-mm segment of silicone or PVC catheter was disinfected in a 70% ethanol bath. The catheter segment was connected to a sterile ICV device (Fig. 1A). A stopper for the catheter was made by cutting 5.0 mm from the tip of a P10 pipette tip, sealing the end with superglue, cauterizing it, and disinfecting it in a 70% ethanol bath (Fig. 1B, C). Next, a Hamilton syringe was disinfected with 70% ethanol, degassed, and used to draw 10 μL of sterile PBS. The prepared ICV device was connected to the Hamilton syringe and degassed. The total dead space in the ICV device, including the connected catheter, was about 2.5 μL. Then the ICV device connected with the Hamilton syringe was set aside on a sterile surface for later use (Fig. 1D).

Figure 1. Preparation of the intracerebroventricular (ICV) device.

(A) A silicone catheter was disinfected in a 70% ethanol bath, cut to 15.0 mm in length, and connected to a sterile ICV device. (B) A P10 pipette tip was cut to 5.0 mm to make a stopper for the silicone catheter. The finer end of the stopper was sealed with superglue and cauterization and the stopper was disinfected in a 70% ethanol bath. (C) The ICV device is shown once plugged by the stopper. (D) A disinfected Hamilton syringe with a 22G needle containing sterile phosphate-buffered saline (PBS) was connected to the ICV device. Flushing with PBS (until 5 μL of PBS was left in the syringe) was used to degas the ICV device.

2.4. Animal preparation and ICV device installation

All experiments were conducted under protocols approved by The University of Texas MD Anderson Cancer Center’s Institutional Animal Care and Use Committee. The protocol we used for ICV device installation was adapted from the protocol described by DeVos et al¹⁰. First, each mouse was anesthetized via continuous administration of gaseous isoflurane. The fur around the head and the anterior back was removed, the skin was disinfected with 10% povidone-iodine, and a 1.0-cm midline linear incision was made between the neck and the occipital bone. Blunt dissection was used to create a subcutaneous pocket in the back. Next, the bregma on the skull was identified and a 26G needle was used to drill a hole at the point 1.1 mm to the right of and 0.5 mm posterior to the bregma¹⁰ The same needle was used to scratch a 2.0 mm × 2.0 mm area on the skull surface around the hole (Fig. 2A). A small drop of superglue was applied to the base of the ICV device (just enough to cover a quarter of the base of the ICV device). Then the ICV device was installed in the predrilled hole, and the surrounding skin was pulled away to avoid its adhesion to any excess superglue (Fig. 2B). The holder of the ICV device was removed with cutting pliers while the base of the ICV device was secured with tweezers. Finally, 5 μL of PBS was infused for 1 minute to confirm the patency of the connection and identify any leakage. Mice used were male due to our focus on cell therapy of the P301S Alzheimer’s Disease model; however, we do not expect significant differences across gender given publications showing that sexual difference in skull landmarks is minimal¹¹

Figure 2. Intracerebroventricular (ICV) implantation.

(A) A linear incision of 1.0 cm was made over the scalp and a 26G needle was used to drill a hole 1.1 mm to the right of and 0.5 mm posterior to the bregma.⁹ (B) The prepared ICV device was installed in the predrilled hole, and 5 μL of phosphate-buffered saline (PBS) was infused to check the patency of the connection. The catheter was closed with the prepared stopper and stored in the subcutaneous pocket on the mouse’s back. (C) The wound was closed with autoclips.

2.5. Infusion of cellular therapy via the ICV device

A Hamilton syringe containing 3 μL of cellular solution of engineered monocytes was connected to the catheter. The cells were infused slowly over 1 minute. Then the syringe was replaced with another Hamilton syringe containing 4 μL of PBS, which was infused slowly over 1 minute to flush the cells toward the ventricle. The syringe was removed, the opening of the silicone catheter was closed with the previously prepared stopper, or the opening of PCV catheter was sealed by cauterization. The catheter was placed in the subcutaneous pocket created in the back, and the incision was closed with 2 autoclips and care was taken to avoid clamping the underlying catheter (Fig. 2C). In a different experiment, the wounds were closed by horizontal mattress suture with a 5–0 nylon thread¹⁰. Finally, the mouse was removed from the surgery station and placed under a heat lamp for postoperative recovery. The wound was checked daily and the clips or the sutures were removed 10–14 days after surgery.

2.6. Recurrent infusion via the ICV device

For recurrent infusions, each mouse was anesthetized with gaseous isoflurane and its skin was disinfected with 10% povidone-iodine. Then, scissors were used to make a 1.0-mm incision approximately 2.0 mm posterior to the tip of the subcutaneous catheter (Fig. 3A), and to gently dissect the subcutaneous tissue in the direction of the catheter. One pair of tweezers was used to locate the incision while another pair was used to hold the skin overlying the catheter. The skin was pulled toward the ICV device to expose the tip of the catheter. Next, a povidone-iodine pad was placed under the catheter and the stopper on the silicone catheter, or the seal of the PVC catheter was removed (Fig. 3B). A Hamilton syringe containing the cellular solution was connected to the catheter and the solution was infused slowly over 1 minute (Fig. 3C). Next, the syringe was replaced with another syringe containing 4 μL of PBS, which was infused slowly over 1 minute to flush the cells toward the ventricle. The stopper was replaced on the silicone catheter, or the opening of the PVC catheter was cauterized before the catheter was placed back into the subcutaneous pocket by gently sliding the skin toward the caudal end. Finally, the incision was closed with an autoclip (Fig. 3D) and the mouse was removed from anesthesia and placed under a heat lamp for postoperative recovery. The mice were humanely euthanized after 8 weekly infusions of cellular therapy.

Figure 3. Recurrent intracerebroventricular (ICV) infusion.

(A) The skin was disinfected, and a 1.0-mm incision was made approximately 2.0 mm posterior to the tip of the subcutaneous catheter. (B) The skin was pulled toward the ICV device to expose the catheter through the incision. The catheter was placed on a sterile pad on the mouse’s back. (C) The stopper was removed and the prepared Hamilton syringe was inserted for infusion. (D) The stopper was replaced, the catheter was stored in the subcutaneous pocket, and the wound was closed with a clip.

2.7. Trypan blue staining of ventricles

In a separate experiment, 20 uL of trypan blue was infused over one minute via an ICV device immediately after its installation in C57BL/6J mice. The mice were humanely euthanized within 5 minutes after the infusion, and their brains were sectioned to assess the extent of the trypan blue staining.

2.8. Imaging of Bioluminescence

We imaged the mice 1 hour, 1 day and 6 days after they received the engineered cells. Mice were imaged by using the IVIS Spectrum in vivo imaging system (PerkinElmer) either 1) 12 minutes after intraperitoneal injection of 3 mg of luciferin (Gold Biotechnology, #LUCK), or 2) 5 minutes after ICV infusion of 150 μg of Luciferin via an ICV device or using a stereotactic device.

2.9. Statistical analysis and illustrations

Kaplan-Meier curves demonstrating time to wound dehiscence between different surgical methods and time to puncture between two materials were plotted using GraphPad Prism, version 8.0. The log-rank test was used to determine statistical significance. P values less than 0.05 were considered significant. The illustrations were created with BioRender.com.

3. Results

Trypan blue staining confirmed that the ventricles are located 1.1 mm lateral to and 0.5 mm posterior to the bregma at a depth of 2.5 to 3.0 mm, as reported by DeVos et al (Fig. 4A)¹⁰. We validated this information and our successful access to the ventricle by injecting Trypan blue via the installed ICV device. Diffuse staining of Trypan blue in the ventricles across brain sections confirmed proper installation (Fig. 4B). In another experiment with repeated treatments, the ICV device remained securely attached to the skull surface during the 8 weekly infusions of cellular therapy (Fig. 4C). These results validate both our protocol for installing ICV devices and the durability of these ICV devices, which we demonstrate at twice the duration of prior methods in a setting where the pumps remained stable and injections could have continued even further.

Figure 4. Validation of the coordinates of the lateral ventricle and confirmation of the durability of the implanted intracerebroventricular (ICV) device.

(A) An ICV device was installed in the predrilled hole 1.1 mm to the right of and 0.5 mm posterior to the bregma. A depth of 2.5 to 3.0 mm below the skull surface was reported to be the optimal depth for lateral ventricle access.¹⁰ (B) Successful implantation was demonstrated by Trypan blue staining of the ventricles of a male C57BL/6J mouse in different coronal sections. (C) In a separate experiment, each male P301S mouse implanted with an ICV device received a weekly infusion of either phosphate-buffered saline (PBS) or engineered cells for 8 consecutive weeks. Mice were humanely euthanized 5 days after the last treatment. The skin covering the skull was removed to show the ICV device.

To validate the utility of the ICV device for the delivery of cellular therapy, we engineered the monocyte cell line PMJ-2PC to overexpress Luciferase. On day 1, we infused 1.0 × 10⁶ engineered cells into the right cerebral ventricle via the ICV device. A mouse infused with the cells transduced with a control vector served as the negative control. The mice were imaged 1 hour after the infusion of cells. Twelve minutes after intraperitoneal injection of 3 mg of Luciferin, in vivo imaging revealed no significant Luciferase activity (Fig. 5A). After waiting for 50 minutes after the injection to avoid signals from the previous dose of Luciferin,¹² we infused 150 μg of Luciferin via the ICV catheter. Imaging 5 minutes after infusion showed significant luminescence in the mice that received the Luciferase-expressing monocytes (Fig. 5B). On day 2, 150 μg of Luciferin was injected into the left cerebral ventricle to confirm that the infused cells were located inside the ventricle. Again, significant signals were observed in the mice that received Luciferase-expressing monocytes 5 minutes after the ICV infusion of Luciferin (Fig. 5C). This finding demonstrates that the cells were successfully delivered into the ventricles and that the ICV installation did not result in detectable extravasation of Luciferin into the ventricles which would signify increased permeability or disruption of vasculature. On day 7, 5 minutes after Luciferin was injected via the ICV device, the mice were imaged again using the IVIS in vivo imaging system. Signals from the mice receiving the Luciferase-expressing cells were dramatically decreased compared to the signals on days 1 and 2 (Fig. 5D). This result was consistent with the previous finding that the number of myeloid cells delivered to the brain diminishes after 1 week.⁷

Figure 5. Kinetics of the intracerebroventricular (ICV) device-infused, engineered monocytes in ventricles.

(A) We infused 1.0 × 10⁶ engineered monocytes with or without Luciferase expression into the right cerebral ventricle of three male C57BL/6J mice via the ICV device on day 1. When the mice were imaged 12 minutes after 3 mg of Luciferin was given intraperitoneally, no significant Luciferase activity was observed. (B) After 50 minutes, 150 μg of Luciferin was given via the ICV device. When the mice were imaged 5 minutes after the infusion, significant luminescence was seen in the mice that had received Luciferase-expressing monocytes. (C) On day 2, 150 μg of Luciferin was given through the left cerebral ventricle. After 5 minutes, the mice were imaged to confirm that the cells were located in the ventricles, and significant luminescence was seen in the mice that had received Luciferase-expressing monocytes. (D) On day 7, 150 μg of Luciferin was given via the ICV device. The mice were imaged again after 5 minutes, and the signals in the mice receiving the Luciferase-expressing cells were dramatically lower than those seen on days 1 and 2.

To reach the cerebral ventricle at a depth of 2.5 to 3.0 mm, four 0.5-mm-thick spacers were needed for Kit 2, resulting in a 2.0-mm elevation of the device above the skull surface. We found that this elevation increased the rate of wound dehiscence in mice in which Kit 2 was used compared to those in which the ICV device that did not require any spacers was used (P < 0.0001) (Fig. 6A). This finding suggests that selecting an ICV device with an optimal needle length is critical to successful infusion and to durable maintenance of the ICV catheter installation – a key advance over prior methods. We further evaluated the wound closure method by comparing sutures that were suggested in the literature¹⁰ with autoclips. We found that the sutures were easily torn, resulting in wound dehiscence. On the other hand, compared with the use of spacers and sutures, the use of no spacers and autoclips offered a significantly longer time to first wound dehiscence (P < 0.001) and providing the durable persistence necessary for long-term cell therapy infusions which was lacking in prior methods (Fig. 6A).

Figure 6. Time to complications for different surgical devices and approaches.

(A) We installed intracerebroventricular (ICV) devices with or without four 0.5-mm thick spacers in 4 to 8-month-old male P301S mice and closed the surgical incisions with either sutures or autoclips. The graph shows the time to the first wound dehiscence. n= 10 for Spacer (+) with suture, n= 12 for Spacer (−) with suture, and n= 7 for Spacer (−) with Auto-clip. (B) We installed ICV devices connected to catheters made of polyvinyl chloride (PVC) or silicone. The graph shows the time to the first catheter-related skin puncture. n=7 for both PVC and silicone groups. Log-rank test, ***p < 0.001, ns: not significant.

We also found that, in addition to the choice of ICV devices and the wound closure method, the choice of catheter material also can affect the risk of surgical complications. Catheters made of PVC are commonly used and are included in the Alzet kits. The PVC catheters became less flexible when they were cut into shorter segments to minimize dead space. In comparison, the catheters made of silicone were more flexible. The time to first skin puncture achieved with PVC catheters was shorter than that achieved with silicone catheters, but this difference was not statistically significant over the 8 week study (Fig. 6B). This trend, however, supported our observation that using catheters made of silicone instead of PVC could help prevent catheters from penetrating the skin, and could further extend patency of the ICV apparatus.

4. Discussion

This study demonstrated the usefulness of a minimally invasive method for long-term, recurrent infusion of cellular therapy using an ICV device. We also improved the surgical complications associated with the method published by DeVos et al.¹⁰. The device and catheter were durable for at least 8 weeks. However, the durability of the apparatus was limited by surgical wound complications. Specifically, the intactness of the skin overlying the ICV device and the tip of the catheter were the primary determinants of apparatus durability and could be affected by the height of the ICV device, the catheter material, and the wound closure method used.

4.1. Height of the ICV device

Alzet Brain Infusion Kits come with a preset needle size and provide spacers for depth adjustment. The Alzet Brain Infusion Kit 2 is more appropriate for cellular therapy because its wider 28G steel tube produces less shearing pressure on cells than does the 30G steel tube in the Alzet Brain Infusion Kit 3.¹³ To use Kit 2 to reach the ventricles at a depth of 2.5 to 3.0 mm, additional spacers were needed, specifically to add an additional 2.0 mm to the overall height of the device above the skull surface. This elevation of the device complicated wound closures because the amount of skin available to cover the device was nearly insufficient. Closing a wound with limited skin generates tension and is detrimental to wound healing.¹⁴ We found that using an ICV device with a customized intracranial needle length precluded the need for spacers and reduced the related surgical complications. This advance was critical is allowing long term maintenance of the ICV device to accommodate repeated cell therapy infusions over the course of months.

4.2. Catheter material

Our early experiments used PVC catheters, whose openings could be shut by cauterization after infusion. In subsequent surgeries, the cauterized ends were removed and, after injection, the catheters were resealed with cauterization. Although the PVC catheters were initially flexible, they became relatively rigid when trimmed multiple times for repeated surgeries. We also found that the PVC catheters pierced the skin in some mice. To prevent this complication, we replaced the PVC catheters with more flexible silicone catheters. The end of a silicone catheter, unlike that of a PVC catheter, cannot be sealed by cauterization, so a stopper was used to close the end of the catheter to prevent communication between the cerebral ventricle and the subcutaneous space. We found that a P10 pipette tip was ideal for this purpose. These tips are inexpensive, readily available in sterile condition in labs, and can be cut to a suitable length. The central lumen also facilitates manipulation of the tip with tweezers. Using silicone catheters with these stoppers, we did not observe any catheter-related skin punctures in 7 mice over 8 weeks. The silicone catheters, like the PVC catheters, remained patent for the duration of the experiment.

4.3. Wound closure methods

DeVos et al.¹⁰ suggested using horizontal mattress sutures with 5–0 nylon thread to close the wound after ICV device installation. However, we found that when these sutures were used, wound dehiscence was common, especially when the mice were housed together. The use of an ICV device without spacers, such as an ICV device with a customized intracranial needle, decreased the skin tension during wound closure. As the demands of covering such a device were easily met with the available skin, it also enabled the use of autoclips to close the wound. The use of autoclips was efficient, precluded the need for prolonged anesthesia, and delayed wound dehiscence. In addition, in our experience, detachment of the clips due to mice pulling or biting them was rare in contrast to the sutures specified in prior methods.

4.4. Limitations

One limitation of the present study is that infusion of other cell type such as cells with different morphologies has not been examined. For example, neural precursor cell transplantation has been studied in various fields such as strokes¹⁵. We believe that our method could be applied to the cells that can be prepared in suspension for infusion. Previous literature indicates that biomechanical force could alter the differentiation of neural precursor cells¹³. It would be ideal to standardize the cellular density for infusion, the needle bore, and the infusion rate while working with these cells.

Our study utilized exclusively male mice for our mouse model, as suggested by the literature for the Alzheimer’s model which is the primary focus of our cell therapy efforts¹⁶. As demonstrated previously, the sexual difference in skull landmarks is minimal¹¹. We would not anticipate a major variation that necessitates a different implantation method. Age was also not determined for its effect on wound dehiscence in this specific surgery. The age of the mice used in our investigation was determined by reviewing published reports of pathology^17,18. A recent study found that young mice heal wounds faster than elderly mice in the first four days, but this advantage fades since day six. Due to the seven-day interval between surgeries, we anticipate that both young and old mice will heal similarly by the time of the next surgery.

5. Conclusion

Thorough preclinical evaluation of the potential benefit of cellular therapies for both brain malignancies and neurodegenerative diseases requires a reliable method for the long-term, repeated infusion of cells into the rodent CNS. The minimally invasive protocol for the long-term, recurrent ICV infusion of cellular therapy describe herein will enable such preclinical investigations of cellular therapy in the CNS that reflect patients’ clinical treatment schedules and allow more realistic assessment of translational potential.

Highlights.

• The method provides durable access to the cerebral ventricles of mice.

• This access enables recurrent intracerebroventricular infusion of cells.

• The procedure is minimally invasive and, unlike prior methods, has few complications even when maintained over weeks to months.

• This method can be used to investigate adoptive cellular therapies for a wide variety of brain pathologies.