Feasibility of an implantable bioreactor for renal cell therapy using silicon nanopore membranes.

The definitive treatment for end-stage renal disease is kidney transplantation, which remains limited by organ availability and post-transplant complications. Alternatively, an implantable bioartificial kidney could address both problems while enhancing the quality and length of patient life. An implantable bioartificial kidney requires a bioreactor containing renal cells to replicate key native cell functions, such as water and solute reabsorption, and metabolic and endocrinologic functions. Here, we report a proof-of-concept implantable bioreactor containing silicon nanopore membranes to offer a level of immunoprotection to human renal epithelial cells. After implantation into pigs without systemic anticoagulation or immunosuppression therapy for 7 days, we show that cells maintain >90% viability and functionality, with normal or elevated transporter gene expression and vitamin D activation. Despite implantation into a xenograft model, we find that cells exhibit minimal damage, and recipient cytokine levels are not suggestive of hyperacute rejection. These initial data confirm the potential feasibility of an implantable bioreactor for renal cell therapy utilizing silicon nanopore membranes.

Silicon membranes for extracorporeal life support: a comparison of design and fabrication methodologies.

Extracorporeal life support is an advanced therapy that circulates blood through an extracorporeal oxygenator, performing gas exchange outside the body. However, its use is limited by severe complications, including bleeding, clotting, and hemolysis. Semiconductor silicon-based membranes have emerged as an alternative to traditional hollow-fiber semipermeable membranes. These membranes offer excellent gas exchange efficiency and the potential to increase hemocompatibility by improving flow dynamics. In this work, we evaluate two next-generation silicon membrane designs, which are intended to be mechanically robust and efficient in gas exchange, while simultaneously reducing fabrication complexity. The "window" design features 10 µm pores on one side and large windows on the back side. The "cavern" design also uses 10 µm pores but contains a network of interconnected buried caverns to distribute the sweep gas from smaller inlet holes. Both designs were shown to be technically viable and able to be reproducibly fabricated. In addition, they both were mechanically robust and withstood 30 psi of transmembrane pressure without breakage or bubbling. At low sweep gas pressures, gas transfer efficiency was similar, with the partial pressure of oxygen in water increasing by 10.7 ± 2.3 mmHg (mean ± standard deviation) and 13.6 ± 1.9 mmHg for the window and cavern membranes, respectively. Gas transfer efficiency was also similar at higher pressures. At 10 psi, oxygen tension increased by 16.8 ± 5.7 mmHg (window) and 18.9 ± 1.3 mmHg (cavern). We conclude that silicon membranes featuring a 10 µm pore size can simplify the fabrication process and improve mechanical robustness while maintaining excellent efficiency.

First Implantation of Silicon Nanopore Membrane Hemofilters.

An implantable hemofilter for the treatment of kidney failure depends critically on the transport characteristics of the membrane and the biocompatibility of the membrane, cartridge, and blood conduits. A novel membrane with slit-shaped pores optimizes the trade-off between permeability and selectivity, enabling implanted therapy. Sustained (3-8) day function of an implanted parallel-plate hemofilter with minimal anticoagulation was achieved by considering biocompatibility at the subnanometer scale of chemical interactions and the millimeter scale of blood fluid dynamics. A total of 400 nm-thick polysilicon flat sheet membranes with 5-8 nm × 2 micron slit-shaped pores were surface-modified with polyethylene glycol. Hemofilter cartridge geometries were refined based on computational fluid dynamics models of blood flow. In an uncontrolled pilot study, silicon filters were implanted in six class A dogs. Cartridges were connected to the cardiovascular system by anastamoses to the aorta and inferior vena cava and filtrate was drained to collection pouches positioned in the peritoneum. Pain medicine and acetylsalicylic acid were administered twice daily until the hemofilters were harvested on postoperative days 3 (n = 2), 4 (n = 2), 5 (n = 1), and 8 (n = 1). No hemofilters were thrombosed. Animals treated for 5 and 8 days had microscopic fractures in the silicon nanopore membranes and 20-50 ml of transudative (albumin sieving coefficient θalb ~ 0.5 - 0.7) fluid in the collection pouches at the time of explant. Shorter experimental durations (3-4 days) resulted in filtration volumes similar to predictions based on mean arterial pressures and membrane hydraulic permeability and (θalb ~ 0.2 - 0.3), similar to preimplantation measurements. In conclusion, a detailed mechanistic and materials science attention to blood-material interactions allows implanted hemofilters to resist thrombosis. Additional testing is needed to determine optimal membrane characteristics and identify limiting factors in long-term implantation.