Effect of Cerebrospinal Fluid on Fibroblasts Concerning Epidural Fibrosis: An In Vitro Study

One of the most common treatments for lumbar disc herniation and other lumbar disorders is lumbar laminectomy. Unfortunately, there may be some unwanted and serious complications with this procedure such as the “failed back surgery syndrome (FBSS)” (1, 2, 3). FBSS is characterized by long term unsatisfactory relief or recurrence symptoms in patients who had laminectomies performed (4).

According to the literature, epidural fibrosis (EF), which emerges as the main cause of FBSS, results from the proliferation of fibroblasts, which is followed by the transformation of fibroblasts to myoblasts and the accumulation of disorganized ECM proteins (4, 5, 6). FBSS is seen in 8-48% of patients that have undergone lumbar disc surgery (6,7). Postoperative epidural fibrosis is an unavoidable part of the healing process following lumbar disc surgery. As high as 25% of the FBSS case are thought to be due to EF. Lack of effective treatment against EF brings the attention back to surgical technique when it comes to scar formation (8). Epidural adhesions make re-exposure of the same operation area technically difficult and dangerous because of the greatly increased risk in nerve injury and dural tears (9).

According to the current literature and practice techniques, different agents are being used to prevent EF formation. Some of these agents are biological barriers such as animal collagen membranes, Gelofoamand there are also different varieties of biological and non-biological agents such as Adcon-L, honey and anti-inflammatory (7). To date there is no single agreed upon treatment method of EF (4). One strategy to reduce EF included the use of mitomycin C, commonly used in cancer treatment, to induce apoptosis in the fibroblasts. Mitomycin achieves this by regulating the miR200b expression and thereby the RhoE expression (10).

Spinal surgery may also have other complications such as cerebrospinal fluid (CSF) fistula and pseudomeningocele, which may occur following an incidental durotomy (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13).

In our clinical practice, we observed that the patients lack EF in the presence of CSF in epidural space. Therefore, we wanted to explore the effect of CSF on fibroblast activity in an attempt to prevent EF.

Epidural fibrosis (EF) was initially discussed in the literature in 1948. It is characterized by fibroblast accumulation, extracellular matrix protein deposits and the inflammation induced distortion of the normal tissue(6,14). TGF-1b and IL-6, inflammatory cytokines, have been suggested to play a role in the development and promotion of EF(4). Fibroblasts produce large amounts of TGF- 1b (6,15).Excessive levels of TGF- 1b stimulates fibroblast proliferation, and eventual differentiation of the myofibroblasts. This also accelerates the deposition of ECM by the myofibroblasts. The ECM proteins include fibronectin which will contribute to the process of fibrosis (6,16). This cascade through TGF- 1b is thought to be the main reason for excessive epidural fibrosis in laminectomized spines (5,6).

Figure 3

Many different methods and substances have been tried and experimented within an effort to prevent EF with varying degrees of success (1,4,6, 7, 8, 9, 10 14,17, 18, 19). Some of these methods include the application of a physical barrier such as Adcon L and polytetrafluoroethylene membrane in an effort to prevent epidural adhesion. However, placement of these barriers and gel foams on the roots and dura post-surgery did not reach the desired effectiveness nor did they improve the magnetic resonance imaging (MRI) findings. The pharmaceutical agents and chemicals employed were not any better in preventing EF. They included many chemicals ranging from immune modulating drugs, metabolic substances, corticosteroids to traditional Chinese medicine. Recently, many agents such as tacrolimus, 10-hydroxycamptothecin, and tamoxifen, suberolyanilide hydroxamic acid, apigenin and emodin has been verified to be effective in inhibiting fibroblast proliferation and reduce epidural fibrosis (20, 21, 22). Gels including ADCON-L, sodium hyaluronate, Gel Amidon Oxyde, high molecular weight hyaluronan, poly(D,L-lactic acid-co-glycolic acid)- poly(ethylene glycol)-poly(D,L-lactic acid-co-glycolic acid) (PLGA-PEG-PL-GA) thermogel were also investigated as promising devices that could reduce epidural scaring and dural adhesion (23). All these agent and gels have shown limited clinical efficacy until now. The clinical application of most of these chemicals are limited due to various reasons such as drug toxicity and adverse effects (9). Perhaps what is also very common regarding these chemicals is that all of them come at a certain monetary cost to the patient but the use of CSF, their own, would be of no monetary consequence.

Cerebrospinal fluid, as the name suggests, is a clear and colourless liquid found around the brain and spinal cord. It is derived from the blood and secreted by choroid plexus (CP) in the ventricles of the brain. Ependymal cells of choroid plexus produce cerebrospinal fluid continuously. The main task of cerebrospinal fluid is to provide mechanical and immunological protection for the central nervous system (24). An average adult is thought to have around 150 ml of CSF, produced at a rate of 0,3-0,4 ml/min (430-530 ml/day) (25,26). The CSF is thought to be 99% water and remaining 1% is made up of proteins, glucose, neurotransmitters and ions. Trace amounts of cell and immunoglobulins can also be found in the CSF (less than 5 cells/ml). The protein content and concentrations vary with different pathologies and from person to person (25, 26, 27).

Diverse molecules such as miRNA, hormones and lipids may even be present in the CSF (25).

In our experimental in vitro study, the strongest difference of cell growth was observed when the 25% and 50% CSF exposed cells are compared to the 100% CSF exposed cells in all the measurement times (24h, 48h and 72h). CSF, in general, contains a higher concentration of sodium, chloride and magnesium and lower concentrations of potassium and calcium when compared to plasma (25). As a result of the lower concentrations of potassium, intracellular potassium in fibroblasts is probably depleted as well. These fibroblasts with the depleted potassium lose their ability for attachment and spreading which are necessary for fibrosis (28). This leads us to hypothesize that the fibroblasts in our study lost their ability to spread and attach probably due to the low potassium concentrations of CSF in the experiment. We do not think that the CSF is having such a toxic effect that the cells are unable to survive at all as the cells were still viable in their flask even after 144 hours of culture (data not shown). The CSF may be able to sustain the cells but through a separate mechanism stop fibroblasts from proliferating.

There are limited amounts of immunoglobulins and negligible amounts of white blood cells in the CSF when compared to plasma and lymph (27,29). After the traumatic injury caused during surgery, acute inflammation occurs through cells such as the PNL, monocytes, macrophages. At some point, fibroblasts enter the scene, eventually leading to end result phenomenon that we call fibrosis. Presence of enough CSF during this inflammation and fibrosis stage, can dilute out the effect and impair fibroblast activity.

CSF has never been studied in any experimental model as possible agent for EF prevention. CSF will have some significant differences and advantages compared to artificial chemicals and chemicals derived from other sources. As it is naturally a body produced material, there will not be an issue of toxicity to the person. Additionally, it will be much cheaper and easier to obtain compared to some of the other chemicals tried. As we will be using the person’s own CSF for themselves we do not foresee any issues with immunogenicity of the biomaterial also. As Table 1 indicated the differences observed using different CSF percentages were very significant whereas the time (in hours) was not significant. This could be taken to mean that the different CSF concentrations are showing significant effect in cell proliferation regardless of the time point. This could be seen in Table 2 as well as the percentage comparisons are almost identical in the different time points. It is very difficult to compare this finding to the current literature as this is the first time the CSF is used to look at fibroblast proliferation. CSF has been shown to affect proliferation and migration in stem cells but not in fibroblasts (30). There are examples of either reduced proliferation with different chemicals or no major difference observed different systems used. However, these systems include animal models, non-comparable setup and/or different chemicals that make it difficult for us to compare our results (8,10,19).

Experiments were done using three CSF samples from three different individuals. Each experiment was performed in quadruplicate to account for variability due to pipetting and other experimental errors. The results obtained with CSF1 and 2 were very similar to each other. Although CSF3 presents a similar patter to 1 and 2 the cell numbers were drastically higher. This could be due to individual variability between the different CSFs. As they come from different individuals, there could be subtle differences in protein/ hormone, miRNA or ionic content which could result in a higher growth potential for the cells. This is one of the possible limitations of our study. Perhaps one potential way to pursue this in the future would be to analyse each CSF for content in detail. When the literature was scanned there were examples of pooling the CSF (31,32). However, we do believe that there is a value to using separate CSFs and catching subtle differences. What way be seen as a limitation could also allow us to do a separate analysis which we would not have been able to do otherwise. We still need to point out that we may not have been able to figure out the growth difference as it may be due to multiple factors all at once.

There were some limitations to the study. To begin with we were only able to do the experiment with CSF from 3 different people. A higher number will definitely lead to more statistically valuable results, but it is important to remember that CSF is difficult to obtain in large amounts. And even with 3 individuals we were able to get statistically significant results. Another way around this problem would be to pool smaller CSF samples from different people to get enough CSF for a complete experiment. However, with that option we may be losing our ability to observe individual variation, which we wanted to see. As mentioned earlier, interpersonal CSF differences and our inability to determine where the variation is coming from is another limitation of our study and we plan to do a full CSF content analysis in future experiments. Another limitation of our study is the use of skin fibroblast cells only. There could be other fibroblast contributions to fibrosis in EP such as the muscle fibroblasts. Testing fibroblasts from different sources to see their response to CSF is a future goal.

It is also important to keep in mind that in vitro experiments are no substitute for the actual conditions in in vivo environments, but they do provide a controlled starting point to investigate new paths.

The molecular mechanism though which the CSF is exerting its effect on the fibroblasts will also need to be determined. This may be achieved through looking at the expression profiles of genes in the apoptotic and proliferative pathways. Another possible molecular analysis could be done through global expression profiling of genes and miRNA in fibroblasts with exposure to varying percentages of CSF. A recent study by Wu. Et al has looked at the effect of mocetionstat and has observed that it exerts its effects through HDAC1 on the AKT/GSK2b pathway to suppress epidural fibrosis (3). A similar molecular profiling for CSF remains a future goal. Another future goal for is to combine CSF with materials such as PLGA that may regulate the release of the CSF over time and lead to a better effect.

In conclusion, adverse effects of low concentrations of potassium levels and dilutional effect of CSF may be a promising solution in the prevention of EF. The exact application method and duration of the application will need to be determined following further in vitro and in vivo experiments, which will include animal experiments (with rats).