Limits in the microscopic understanding of chorioamnionitis

14 Nov

(I have been pondering the patterns of chorioamnionitis for the placenta section on the web site. I will post a rough draft without images so that those interested can find the details and references that these musings are based on. http://www.pediatricperinatalpathology.com)
As pathologists we are used to adjusting for the fact that we must visualize three dimensions from simple two dimensional planar sections. For the non-pathologist, think of recognizing the difference between cross sections of spheres versus cylinders to get a perspective on this every day difficulty. Even more difficult is trying to infer information about the fourth dimension, time. Our histological sample represents a fixed point in time, but we try to analyze dynamically, for example an early as opposed to a late stage of a process.
One way pathologist overcome this problem is to use the concept of staging. If a large series of cases always shows the same progression of a process, for example if lymphatic invasion of a tumor is always associated with local invasion, but local invasion can be present without lymphatic invasion then it is logical then to infer that local invasion precedes lymphatic invasion. This inference may be correct, but still not tell us about timing in that getting from mucosal invasion to lymphatic invasion may not require the same interval in every patient. For example, if the change in invasion requires a second mutation in the tumor, this might be a random event that could take place at any time. At the extreme, some of the tumors that seemed to fit the pattern might never develop lymphatic invasion. So staging is not the equivalent of timing. There are several systems for histologic staging of chorioamnionitis.
Another way that pathologists estimate the timing of histological events is to compare the findings to those in experimental models such as the healing of a clean scalpel wound in an animal or the organization of an induced thrombus. There are limitations in applying experimental conditions to human pathology but it does provide a direct estimate of timing if conditions are applicable. Human fetal membranes have been exposed to neutrophils and bacteria in vitro.
Yet another way is to grade a progression of pathological changes is to use a well-defined clinical starting point. The classic study of Mallory and Weiss timing the changes following myocardial infarction is a well-known (by pathologists) example. An ingenious study by Genest and colleagues used this approach to timing the duration of retention of a deceased fetus in the uterus by creating windows for each case bounded by the last known time the fetus was alive and the first time that the fetus was documented to be dead. There is no such starting point for chorioamnionitis but timing of labor and ruptured membranes are potential points to use for comparisons if confounding variables were controlled.
All of these techniques have significant limitations, for example wound repair is often not sterile, and the Genest timing could be influenced by necrosis of organs prior to fetal death or to postmortem storage at warm temperature. Still, such attempts at timing are often useful in understanding the progression of histologic change with disease or death.
These approaches to timing fall short in trying to understand the progression of chorioamnionitis. I have been on the opposing side in medico-legal consultation in which an expert has opined that a certain stage of chorioamnionitis implies that the infection has been present for a certain period of time. My contention has been that we can not estimate the duration of infection, nor estimate the timing of the inflammation. For the first, we do not how long bacteria have to be present or in what numbers to induce a generalized inflammatory response, so we can not estimate when infection started from the stage of inflammation. I also argue that there are too many variables to time the inflammation from the staging in that the species of microorganisms, individual variations in maternal or in fetal inflammatory responsiveness, and variance in the numbers of microorganisms might all influence the rate of response. These are obvious objections, but they don’t get to the heart of some questions that staging of chorioamnionitis poses.
In chorioamnionitis, neutrophils are commonly seen piled up at certain locations, rather than showing a smooth progression. This makes staging easier since the neutrophils are likely to be confined to a certain area, such as the subchorionic intervillous fibrinoid, at the interface of the chorionic epithelium and chorionic connective tissue, just beneath the amnion basement membrane. In the one experimental in vitro study that I found the neutrophils cross the membranes quickly at least going backward (the experimental neutrophils were applied to the amnion side). It is unlikely that we just happen to capture the histology in this limited number of stages during a continuous process. The neutrophils are likely staying where they are because the chemotactic stimulus fixes them there. Since at later stages neutrophils have transversed these stopping points in large numbers, it is less likely they accumulate because of a physical barrier to their progression.
One reason that neutrophils appear to stop in sections is that they encounter microorganisms. This is occasionally evident in subacute necrotizing funisitis and in the membranes with filamentous organisms in the chorion connective tissue or cocci on the surface of the amnion epithelium. Since microorganisms are not seen histologically in most cases of chorioamnionitis, it is not possible to know how often the neutrophils are stopping because they have met the enemy. It is possible that some potent chemotactic signals are localized in a way that causes neutrophils to accumulate focally. My problem is that I can not tell looking at the slides where the organisms are, which cells are making the important local chemotactic signals to attract neutrophils, or how such chemo-attractants are diffusing their way through the tissue. It is this ground level detail, as would be perceived by the neutrophil as they move through the membranes that is frustratingly not apparent to the human microscopist.
There is another important problem related to timing of chorioamnionitis, and that is why we do not see the ascending nature of the infection. In general, the inflammation shows a relatively homogenous stage and intensity in any one placenta. The simplest explanation is that both microorganisms and chemotactic inflammatory chemicals are mixed homogenously in the amniotic fluid and distributed evenly across the membranes and placental surface. Does this concept apply following rupture of membranes? Perhaps there is enough re-accumulated fluid to accomplish the same task. Rarely in preterm premature rupture of membranes in which we orthogonally sampled 4 directions of membranes from the ruptured membrane edge, there was inflammation only along the rupture margin. One possible inference from the even distribution of inflammation is that chorioamnionitis may start with very few organisms that need time to proliferate in the fluid or membranes before eliciting histologic chorioamnionitis. These few organisms have time to spread out as they reproduce. The paucity of species of microorganisms cultured from any one patient also supports this hypothesis that the original inoculum might be very small.
I think that our inability to deeply understand the relationship of chorioamnionitis and preterm labor justifies further thought about the microscopic diffusion of inflammation and infection in chorioamnionitis. There are many potential ways to study the problem, from experimental models in animals or in vitro membranes, to sequenced clinical biopsies perhaps via an intrauterine pressure catheter (taken very carefully at microscopic bite sizes). Membranes taken immediately following delivery could be optimally processed for in situ PCR of microorganisms or for immuno-localization of chemotactic chemicals.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: