The two-hit hypothesis has been the prevailing paradigm in ADPKD research for decades, but it has failed to guide effective treatments for the disease(24, 25). We challenged this hypothesis by examining the three-dimensional morphology of renal cysts. Surprisingly, in rapid-onset model, we found that the cysts in the rapid model were not clustered like grapes on a certain renal tubule(26), but rather the entire renal tubule elongated like cords. This suggests that cyst formation is not due to random mutations in some cells that become malignant and proliferate, but rather to obstruction of a segment of the renal tubule that causes accumulation of primary urine and forced expansion near the Bowman’s capsule. This explanation is consistent with the observation that everolimus, a drug that inhibits cell proliferation, can slow down kidney volume increase in ADPKD patients, but cannot improve kidney function(24, 25, 27). If cell proliferation was the main cause of cyst formation, then reducing kidney volume should preserve kidney function. However, this is not the case, implying that obstruction of renal tubules is the primary mechanism of cystogenesis and kidney dysfunction in ADPKD.
In a previous study using an ADPKD mice model, we found that inhibiting cell proliferation could reduce cyst size, but not cyst number or rescuing kidney function(27). This was contrary to the expectation based on the proliferation hypothesis, which predicts that blocking cell proliferation should prevent new cysts from forming. However, we observed that the treated mice had the same number of cysts as the control mice, but each cyst was thinner. This can be explained by the obstruction hypothesis: when a segment of the renal tubule is blocked, the epithelial cells proliferate under pressure and form a cyst. When proliferation is blocked, the cyst becomes thinner, but the obstruction remains, so the number of cysts does not change. Moreover, this treatment does not improve kidney function, because the obstructed renal tubules cannot perform their normal functions.
we show cord-like structures in the tubular lumen, composed of epithelial cells and basement membrane. We have previously seen similar structures in the kidneys of ADPKD mice(28), but we did not pay much attention to them, because they seemed irrelevant to the disease process, or the observed interference could be attributed to artifacts during the pathological sectioning process. However, since tissue clearing followed by 3D imaging is a non-destructive imaging technique, it indicates that this is indeed a naturally occurring structure(23). Here we propose a hypothesis that the process of tubular obstruction in the kidney may resemble the process of spider web formation.
Indirect evidence for this hypothesis can be found in the study by Dong K et al (20), where it was discovered that renal tubular epithelial cells undergo a transformation similar to EMT upon loss of PKD1. These cells change from a cuboidal epithelial morphology to an elongated flattened form. Additionally, the expression of laminin γ1 protein on the basement membrane of the renal tubules is significantly reduced. Laminin γ1 is an adhesive protein that helps anchor cells to the basement membrane, allowing epithelial cells to detach and interconnect with each other. This process resembles the formation of a spider web. In the spider web analogy, the spider needs to find a way to attach a thread between two branches, and then use this thread as a scaffold to weave a web. Moreover, due to the morphological changes in renal tubular epithelial cells, their ability to maintain their own shape is weakened, making renal tubule more susceptible to compression by surrounding tissues. In the obstruction hypothesis, the renal tubular epithelial cells undergoing EMT in the lumen of the renal tubules are such threads. If they connect the two sides of the tubular lumen, and then cell growth blocks the lumen, an obstruction is formed; otherwise, if the urine flow is strong enough, these threads fail to connect, and this tubule is spared from cyst formation.
Our model also provides a good explanation for why restoration of PKD1 or PKD2 gene expression can reverse the disease process. Our chronic disease 3D model results indicate that although all cysts may appear to be similar on the surface, one type of cyst is still connected to the renal pelvis through its collecting duct opening. This particular cyst regains the ability of renal tubular epithelial cells to sense urine flow, leading to a reoccurrence of mesenchymal- epithelial transformation and subsequent disappearance of "cobwebs". The duct undergoes remodeling, gradually becoming more patent, thus achieving partial morphological restoration. On the other hand, another type of cyst that does not communicate with the renal pelvis cannot return to normal even if the renal tubular epithelial cells regain their ability to sense urine flow, as the stagnant fluid within the cyst hinders the restoration process. This is described in Dong et al’s article as “tubule dilation with protein casts”.(20) This phenomenon can explain both partial recovery and permanent loss of renal function.
Tolvaptan, a vasopressin V2 receptor antagonist, is the only approved drug that can effectively slow down kidney volume growth and preserve renal function in ADPKD patients(29, 30). This is remarkable, considering that all anti-proliferative treatments have failed to achieve the same results. Before tolvaptan, the management of ADPKD was limited to symptomatic treatments, such as blood pressure control, cyst complications treatment, pain control, and lifestyle changes (salt intake restriction and increased water intake). However, there is no strong evidence that tolvaptan works by inhibiting cyst cell proliferation. In fact, it would be surprising if it did, since it is a diuretic that has a much better clinical efficacy than sirolimus and everolimus, which are anti-proliferative drugs.
Based on the results of this study and the aforementioned hypothesis, a more plausible explanation is that during the development of the kidney, some tubules may not successfully connect to functional glomeruli, resulting in ineffective dead spaces. The primary urine in these dead space tubules is almost stagnant, and no water flow is formed. This low flow rate is sensed by the cilia on the tubular epithelial cells, activating some certain genes, and initiating a mechanism to remove this segment of the tubule. The specific action is to grow a cord-like tissue composed of epithelial cells and basement membrane into the lumen, which is the “spider silk” we see. In the stagnant liquid, this section of “spider silk” easily adheres to the opposite side of the tubular lumen, and continues to proliferate, blocking this segment of the lumen, and repeating this process until the lumen is completely filled. This may be a self-correcting mechanism in kidney development.
In previous studies, some chronic-onset rodent models of ADPKD also showed tubular cysts in their kidneys(26), but this structure is incompatible with the proliferation hypothesis and has been consistently ignored. We observed this phenomenon in the chronic-onset murine kidney, where occlusions formed by cellular and proteinaceous deposits were found within completely obstructed renal tubules. After the PKD1/2 gene is mutated, ciliary sensitivity is reduced, and areas where the original urine flow rate is not high enough are mistakenly regarded as dead spaces. The result of blocking these tubules is the formation of obstructive cysts. In this case, the use of tolvaptan at this time to increase the flow rate of primary urine in the tubules can prevent some tubules from being blocked. On one hand, the increased urine flow can be detected by damaged cilia, and on the other hand, it can flush out obstructive substances from the lumen. This is analogous to how spiders have more difficulty weaving webs between branches in strong winds. Similarly, this theory can explain why the condition of some patients suddenly accelerates in progression. When ADPKD develops to a certain stage, due to the compression of the cysts, the flow rate of primary urine in the surrounding tubules decreases, and the blocking mechanism is activated in these tubules. In this way, new cysts are constantly formed around the old cysts, and the condition collapses like dominoes. This can also explain why limiting salt intake and increasing water intake have some therapeutic effects in clinical practice: both of these interventions increase urine volume as well.
The obstruction of the tubules leads to cyst formation, as the primary urine produced by the upstream glomeruli increases the pressure in the dilated tubules. In a short time, the hydrostatic pressure makes the cysts spherical. If the researchers do not catch this brief window of time after the obstruction occurs, they will miss the cord-like cysts, and only see spherical cysts clustered like grapes. This is what has been observed in late-stage ADPKD patients for centuries. Although the three-dimensional structure we observed supports the obstruction hypothesis, which can account for the failure of sirolimus treatment and the success of increased water intake and tolvaptan, we still need to find out how PKD1/PKD2 gene defects cause abnormal obstruction, in order to develop more effective therapies for ADPKD from the root cause.
In conclusion, our study revealed a cord-like three-dimensional structure of cysts formation in Rapid-Onset and Chronic-Onset models, supporting the obstruction hypothesis rather than the proliferation hypothesis. Moreover, we observed the cell bridge blocking the tubular lumen, providing evidence for the obstruction mechanism. The obstruction hypothesis is better able to explain the current known basic research and clinical experimental results than the proliferation hypothesis. However, the link between PKD1/PKD2 gene defects and obstruction, and the potential targets for intervention, remain to be further investigated.