Lung transplantation (LT) is a therapeutic alternative for some terminal respiratory diseases. It is currently indicated in advanced pulmonary emphysema, cystic fibrosis, diffuse interstitial disease and pulmonary hypertension. The most widely used technique is bilateral lung transplantation, followed by single lung transplantation. The importance of improving certain aspects, such as preservation, ischemia time, ischemia reperfusion (IR) injury, acute rejection (AR) and chronic rejection (CR) makes it necessary to develop adequate experimental LT models.1

In the beginning, the most commonly used experimental animal model was dog,2,3 although the ethical and economic problems made researchers look for alternative models, such as rabbits4 and rats. In 1971, Asimacopoulos5 started experimental LT in rats, and in 1982 Wildevuur6 completed the technique successfully, although with difficulties for performing the vascular and bronchial anastomoses with sutures.

The complexity of the microsurgical technique in rats improved in 1989 with the initial technique described by Mizuta.7 It was the first time that Teflon cuffs were used for performing the anastomoses. To create these, polyethylene intravenous catheters were used, cut into 2-mm cylinders. Initially, Mizuta performed the vascular anastomoses in the receptor with cuffs, while the bronchial anastomosis was done with sutures. Later, Mizuta8 and Reis9 reported doing all the anastomoses with cuffs. The modification in the technique significantly reduced the duration of the procedure and post-op mortality, which gave rise to the introduction of a new model for studying LT and its complications.

Since then, some authors10,11 have described modifications in the surgical technique for studying different aspects of LT and improving experimental conditions, but without including important aspects related with clinical LT, such as brain death (BD), retrograde perfusion in the donor or the start of ischemia time in these models.

In our study, we present a new anesthesia protocol and some modifications in the preservation and lung implantation technique in rats with the purpose of reproducing what is currently done humans. Our objective is to reduce possible variations in the surgical technique for the interpretation and extrapolation of the results.

Mean surgical time was 48.5±2.0min for the extraction and 59.2±4.2 for the implantation. Mean time for the complete procedure was 107±6.2min.

No post-op infections were observed. Four of the recipient animals (two from the 7-day group and two from the 90-day group) presented tachypnea and hemoptysis due to severe edema and died within the first 48h. One animal presented tachypnea for the first 24h. The other animals presented no clinical problems, and the survival rate was 80% (Table 1).

In the histologic study, signs of IR injury were observed after 7 days in all the transplanted lungs (n=8): mild (n=5), moderate (n=2) and severe (n=1); signs of AR were seen in 6 animals (A1=3 and A2=3). After 90 days, in all the animals (n=8) a fibrotic lung stump was observed, and microscopically we viewed bronchial dilation with areas of vasculitis and fibrosis accompanied by severe AR (A4), indicating the presence of active CR. No histological changes were observed in the non-transplanted lung. The inspection of the cuffs showed in all animals adequate bronchial and vascular lumen without signs of stenosis or thrombosis.

Chest radiography showed adequate lung expansion with an increase in density of the transplanted lung related to the presence of IR injury after 7 days and CR after 90. One animal presented mild left pneumothorax and another presented contralateral pneumothorax immediately after surgery, with no clinical repercussions.

LT is still a challenge for patients and medical transplantation groups, and there are still important areas requiring basic and clinical research. To face these challenges, the American National Heart, Lung, and Blood Institute14 (NHLBI) recommends considering the following points: (a) to identify essential biomarkers, based on proteomic and genomic studies that can predict the presence of CR; (b) to develop relevant animal models that better reproduce human diseases in order to research the pathogenesis of lung injury after transplantation; and (c) to support and promote research covering the range of graft dysfunctions, including CR.

In this direction, following the recommendations of the NHLBI, the experimental model that we have developed in this study is based on previous reports of cuff techniques7–9 with modifications in the preservation and implantation techniques such as the induction of BD, control during the ischemia time and retrograde perfusion in the donor and controlled sequential reperfusion in the lung of the recipient, in addition to an improvement in the percentage of survival of 80% compared with other transplantation models in rats.11 This is an attempt at developing animal models that reflect what happens in human diseases according to the aforementioned international recommendations.

The organs used for transplantation come mainly from BD donors.15 The majority of the donors suffer considerable damage in the central nervous system due to subarachnoid hemorrhage or traumatic brain injury before being declared BD.16 BD induces autonomic and hemodynamic alterations as well as alterations in the neurohormonal state of the donor17,18 that can induce lung edema and intra-alveolar hemorrhage and should be taken into account in experimental LT models in order to simulate a model that is closer to reality.

Different animal models have been described that evaluate the effects caused in donated organs due to either rapid-onset or gradual-onset BD13,16,18,19 used in transplant studies for kidneys,20 liver21 and very rarely in LT22 (Table 2). In our case, we have used the slow BD model, with a gradual increase of the intracranial pressure in order to reduce the hemodynamic instability of the donor13 and the subsequent degree of lung edema, which in the case of rats is aggravated by the lability of the pulmonary vascular permeability. In fact, in donor BD rats, a severe pulmonary edema is frequently produced more than 2h after the induction of BD; this is why we do the LT with 2h of ischemia time.22

In our model, orotracheal intubation with an Abbocath® 14G minimizes peritracheal air leak and ensures better ventilation of the rat during surgery, as there are no endotracheal tubes for rats. The removal of the donor lung by mid-sternotomy,9,23 unlike Mizuta,7 facilitates the extraction of the thoracic organs and minimizes any possible injury. Similar to transplantation in humans, the clamping of the aorta gives us precise control of the duration of the ischemia, which is not controlled in other proposed models.7,9 This also acts as a mechanical factor favoring cardiac arrest, similar to what happens in humans. In our model, as in that by Reis,9 we prefer 16G cuffs in order to minimize the risk of tearing the hilar structures, especially the pulmonary vein. The reperfusion and permeability have been shown to be good; therefore, we consider them to be adequate for animals between 300 and 400g.

An important part of the extraction process in humans is the anterograde and retrograde organ preservation. During extraction, the flow volume and rate should guarantee homogenous perfusion of the lung. This should be carried out with the lungs ventilated in order to avoid areas of atelectasis and poor perfusion.24 In our model, we have reached optimal perfusion with an Abbocath® 16G using cold low-potassium dextran solution, which is currently the most widely used in clinical practice. The introduction of retrograde perfusion, which has not been reported in other rat models, is important because anterograde perfusion has been shown to be incomplete as it obviates the bronchial circulation.25 Retrograde preservation can be safely done with lower pressure and volume than those used in anterograde perfusion due to the pulmonary vascular lability of this type of animals. This makes our model more similar to what happens in clinical practice (Table 3).

In our case, the extraction of the cardiopulmonary block is similar to what is done in humans, unlike what is described in other experimental models where only the left lung is extracted. This enables us to maintain the ventilation of the lung with FiO2 0.21 and PEEP 2cmH2O24,26 until the ligature of the trachea and extraction of both lungs semi-inflated to 50% of their total capacity, which is achieved with an air pressure of 12–15cmH2O. The dissection of the left hilum with a microscope minimizes the risk of vascular injury and enables the vessels and the bronchus to be cut as long as possible for the placement of cuffs and the later transplantation. The clamping of the left main bronchus before its transection at the carina allows us to maintain the lung ventilated during its preservation prior to implantation.27

The anesthesia protocol, especially in the receptor, allows for the animal to be awakened at any moment without it being necessary to metabolize the anesthesia, add extra doses or use inhaled anesthesia. The effects of medetomidine can be easily reverted with its antagonist, atipamezole.28 In our case, we associate it with ketamine in order to achieve proper sedation, relaxation and hypnosis. In addition, the administration of atropine diminishes the secretions of the bronchial tree and has palliative effects on the secondary effects caused by medetomidine.29 Due to the effect of the anesthesia, hypothermia may appear, which is controlled with an electric blanket under the animal; if it is not corrected, it may cause death. The administration of these three drugs has constituted in our experience a rapid, efficient and safe anesthesia protocol for LT in these animals.

The approach of the thoracic cavity through the fourth intercostal space in the recipient (unlike other authors who use the fifth intercostal space9) allows us to better manipulate the pulmonary hilum for its dissection. In addition, the transection of the pulmonary vessels and the main bronchus as far as possible from the heart and on its anterior side,7 without completely cutting the structure, make the anastomoses much easier. Similarly, using long ligatures behind the microclamps makes it easier to introduce the cuff and close the anastomosis, unlike what is described by Reis.9

Controlled sequential reperfusion24,30 is a critical step in human transplantation, but it has not been described in other proposed experimental models. In our model, we carried out anterograde reperfusion with the ventilated lung, enabling a homogenous distribution of the blood flow of the lung with flushing of the preservation solution out the vein, similar to what happens during transplantations in clinical practice. In addition, it avoids the introduction of massive quantities of preservation solution within the vascular venous flow along with an increase in the lung capillary pressure and a subsequent pulmonary edema unrelated with IR injury, which may cause artifacts in the results (Table 4).

Thus, we conclude that, with the modifications that we have described, the LT technique with cuffs in rats more closely resembles what happens in human clinical practice. This could facilitate the interpretation and clinical extrapolation of the results obtained by working with these experimental models, which are essential in LT research.

This project has been funded by the Spanish Health-Care Research Fund (Fondo de Investigación Sanitaria Español-FIS 07/01145) and by the Fundación MM in Madrid. Dr. Norberto Santana Rodríguez (INT 07/172) and Dr. Bernardino Clavo (INT 07/030) have worked under the Research Intensification Program (Programa de Intensificación de la Investigación-I3SNS) at the Carlos III Health Institute (Instituto de Salud Carlos III). This study has served as a foundation and continues thanks to a grant from the Carlos III Health Institute (FIS PI 10/01485).