Learning interpretable causal networks from very large datasets, application to 400,000 medical records of breast cancer patients.

Discovering causal effects is at the core of scientific investigation but remains challenging when only observational data are available. In practice, causal networks are difficult to learn and interpret, and limited to relatively small datasets. We report a more reliable and scalable causal discovery method (iMIIC), based on a general mutual information supremum principle, which greatly improves the precision of inferred causal relations while distinguishing genuine causes from putative and latent causal effects. We showcase iMIIC on synthetic and real-world healthcare data from 396,179 breast cancer patients from the US Surveillance, Epidemiology, and End Results program. More than 90% of predicted causal effects appear correct, while the remaining unexpected direct and indirect causal effects can be interpreted in terms of diagnostic procedures, therapeutic timing, patient preference or socio-economic disparity. iMIIC's unique capabilities open up new avenues to discover reliable and interpretable causal networks across a range of research fields.

Second Generation of Tissue-Engineered Ligament Substitutes for Torn ACL Replacement: Adaptations for Clinical Applications.

The anterior cruciate ligament (ACL) of the knee joint is one of the strongest ligaments of the body and is often the target of traumatic injuries. Unfortunately, its healing potential is limited, and the surgical options for its replacement are frequently associated with clinical issues. A bioengineered ACL (bACL) was developed using a collagen matrix, seeded with autologous cells and successfully grafted and integrated into goat knee joints. We hypothesize that, in order to reduce the cost and simplify the model, an acellular bACL can be used as a substitute for a torn ACL, and bone plugs can be replaced by endobuttons to fix the bACL in situ. First, acellular bACLs were successfully grafted in the goat model with 18% recovery of ultimate tensile strength 6 months after implantation (94 N/mm2 vs. 520). Second, a bACL with endobuttons was produced and tested in an exvivo bovine knee model. The natural collagen scaffold of the bACL contributes to supporting host cell migration, growth and differentiation in situ post-implantation. Bone plugs were replaced by endobuttons to design a second generation of bACLs that offer more versatility as biocompatible grafts for torn ACL replacement in humans. A robust collagen bACL will allow solving therapeutic issues currently encountered by orthopedic surgeons such as donor-site morbidity, graft failure and post-traumatic osteoarthritis.

Enhanced secretion of TIMP-1 by human hypertrophic scar keratinocytes could contribute to fibrosis.

Hypertrophic scars are a pathological process characterized by an excessive deposition of extracellular matrix components. Using a tissue-engineered reconstructed human skin (RHS) method, we previously reported that pathological keratinocytes induce formation of a fibrotic dermal matrix. We further investigated keratinocyte action using conditioned media. Results showed that conditioned media induce a similar action on dermal thickness similar to when an epidermis is present. Using a two-dimensional electrophoresis technique, we then compared conditioned media from normal or hypertrophic scar keratinocytes and determined that TIMP-1 was increased in conditioned media from hypertrophic scar keratinocytes. This differential profile was confirmed using ELISA, assaying TIMP-1 presence on media from monolayer cultured keratinocytes and from RHS. The dermal matrix of these RHS was recreated using mesenchymal cells from three different origins (skin, wound and hypertrophic scar). The effect of increased TIMP-1 levels on dermal fibrosis was also validated independently from the mesenchymal cell origin. Immunodetection of TIMP-1 showed that this protein was increased in the epidermis of hypertrophic scar biopsies. The findings of this study represent an important advance in understanding the role of keratinocytes as a direct potent modulator for matrix degradation and scar tissue remodeling, possibly through inactivation of MMPs.

New ligament healing model based on tissue-engineered collagen scaffolds.

The anterior cruciate ligament (ACL) is often the target of knee trauma. This ligament does not heal very well, leading to joint instability. Long-term instability of the knee can lead to early arthritis and loss of function. To develop efficient strategies to stimulate posttraumatic ACL regeneration in vivo, a good healing model is needed in vitro. Such a model must remain as simple as possible, but should include key features to provide relevant answers to precise questions about the clinical problem addressed. Here, we report tissue-engineered type I collagen scaffolds developed to establish an ACL healing model in vitro and a potential ACL substitute in vivo. Such scaffolds were used to evaluate ACL cell growth, migration, and the capacity to synthesize and assemble collagen fibers for up to 40 days in vitro and up to 180 days in vivo. They were anchored with two bone plugs to allow their static stretching in culture and to facilitate their surgical implantation in knee joints. Our results have shown that living ACL fibroblasts can attach, migrate, and colonize this type of scaffold. In vitro, the cells populated the scaffolds and expressed mRNAs coding for the prolyl-4-hydroxylase, involved in collagen fibers' assembly. In vivo, acellular implants were vascularized and populated with caprine cells that migrated from the osseous insertions toward the center of the grafts. This model is a very good tool to study ACL repair and identify the factors that could accelerate its healing postsurgery.