More in vivo experimentation is needed in cardiovascular physiology

W. Gil Wier

in this perspective, I present the view that cardiovascular physiologists may do well to reinvent some of our old in vivo experiments to gain a deeper and more useful understanding of physiology and disease. Certainly, many physiological and pathophysiological phenomena were discovered through classic in vivo experiments on cardiovascular function (exemplified particularly by the work of A. C. Guyton and his colleagues in the 1970s), but methodological limitations at that time precluded the elucidation of the underlying molecular mechanisms. During the ensuing four decades, new methodologies and the extensive use of isolated tissues, isolated cells, and cell lines led to the discovery of a myriad of molecular mechanisms that undoubtedly contribute to cardiovascular regulation. Nevertheless, such studies do not necessarily inform us about the relative contributions of these mechanisms in vivo. A case in point: in clinical trials (perhaps the ultimate in vivo experiment) the approval success rate for new cardiovascular drugs is only 8.7% (3). One frequent reason for the poor success is unanticipated complexity of the intact organism. Thus a new, more insightful, in vivo experimental approach is needed for cardiovascular physiologists to elucidate normal physiological and pathophysiological mechanisms.

Many new approaches are now possible, but one with exceptional potential in my opinion is the visualization and quantification of signal transduction within cells in the living, and even conscious, animal. This could be considered a reinvention of classic intravital imaging, extended to the molecular level. But, is this really possible? The answer, quite clearly, is “yes,” and our neuroscientist colleagues are showing the way. For example, calcium signaling is now being observed through a cranial window in cerebral neurons of conscious, behaving mice (5), thus providing a means to associate populations of cells, individual cells, and even parts of cells with behavior, a phenomenon that only exists in living, conscious animals. Recently, in this journal, we reported completely noninvasive, quantitative, two-photon imaging of Ca2+ signaling in individual smooth muscle cells of arterioles of (anesthetized) hypertensive optical biosensor mice (7). As a further example, two-photon microscopy (1) has now made it possible to image blood flow in cerebral blood vessels, noninvasively, through the thinned-skulls of anesthetized or conscious mice (4) and to cause thrombi and strokes in the cerebral circulation through optical means (10). It is the fortuitous confluence of genetic engineering (to produce optical Ca2+-biosensor mice) and two-photon microscopy (which enables high-resolution imaging deep in the brain) that makes this a reality. These techniques make possible serial imaging of an individual over days. This is significant, since it allows longitudinal studies of disease states or pathology with the important improvements in statistical power and reduction in the number of animals used. Although the studies cited above were on readily accessible cerebral blood vessels or subcutaneous arterioles of the mouse ear, these techniques have also been used to visualize [Ca2+] and other signaling molecules in exteriorized kidney (9). Visualization at the molecular level has come to the intact animal. It seems clear that new realms of investigation in the areas of neurovascular coupling, the physiology and pathophysiology of blood flow, stroke, and hypertension have been opened up.

An uncritical adoption of such in vivo experimentation is not likely to be helpful. Thus it is important to define the expectations for in vivo experiments and to compare them with those for ex vivo experiments. As alluded to above, the key point is that ex vivo systems can tell us what is possible at the molecular level and in vivo experiments can tell us what, among the many possibilities, actually happens in the living animal. Two examples, that of RhoA signaling (which activates Rho-Kinase) and of Ca2+ signaling in vascular smooth muscle, will serve to illustrate the point. Both are, obviously, key players in hypertension. Much is known about the RhoA/Rho-kinase signaling pathway and its many potential roles in normal physiology and hypertension. RhoA is regulated by several GEFs, including LARG, PDZRhoGEF, p63RhoGEF, and p115RhoGEF. These GEFs are coupled to both Gαq/11 and Gα12/13 G proteins, which, in turn, couple to receptors for endothelin-1, thromboxane, angiotensin II and norepinephrine, and others. Downstream, Rho-kinase modulates myosin light chain phosphatase (to name just one effector), which is involved in cytokinesis, membrane trafficking, cell migration, smooth muscle contraction, and vascular remodeling. Each step of the Rho-kinase signaling pathway is localized; activation of Rho-kinase by RhoA occurs at the cell membrane, and Rho-kinase effectors are spatially localized. Definition of the regulatory pathway, the many effectors, and the location information has required study of isolated cells or tissues; only in that experimental milieu is it possible to manipulate the pathway to the extent required to define it at the molecular level. But which receptors, GEFs, and effectors are most important in hypertension? Since hypertension, like behavior, exists only in the living animal, in vivo experiments seem almost mandated. Now, some of the same imaging techniques that revealed the details of Rho-kinase signaling in isolated cells can be used in living hypertensive animals. If an optical biosensor animal for Rho-kinase activity were to be created (similar to those existing now for Ca2+), we could expect to learn exactly in what cells, and where in the cells, Rho-kinase is being activated and how it is different in hypertension.

Similar to the case for RhoA signaling, we know already that vascular smooth muscle cells generate an extensive array of spatially localized changes in [Ca2+], all with distinct molecular mechanisms: asynchronous propagating Ca2+ waves, synchronous [Ca2+] oscillations (associated with oscillatory vasomotion), Ca2+ sparks, sparklets, puffs, junctional Ca2+ transients, and flashes. But how are all these possibilities integrated in blood vessels in the animal where autonomic nervous system activity is ongoing, as are endothelial signals and many others? We have already obtained evidence that the naïve predictions one might make about artery Ca2+ signaling in vivo may not be correct. For example, in vivo, we (6) find that asynchronous propagating Ca2+ waves occur at far lower frequency than reported for isolated arteries. If this is indeed the case, then it means that certain molecules, such as inositol 1,4,5-trisphosphate receptors, may have different roles physiologically than might be inferred from the ex vivo studies.

Optical biosensor mice are key to reinvented intravital imaging. For cardiovascular research, the existing optical biosensor animals include those that express an exogenous Förster resonance energy transfer (FRET)-type, Ca2+/calmodulin biosensor based on smooth muscle myosin light chain kinase and those that express a single fluorescent protein, chromophore modulating type of Ca2+/calmodulin sensor GCaMP2. The further development of the optical biosensor mice that use FRET-based biosensors (11) is particularly needed, since these can provide quantification through the ratioing of acceptor and donor fluorescence emission and other techniques. FRET biosensors of the integrated or intramolecular type, in which both acceptor and donor moiety are present in the same molecule, are most advantageous. Ideally, optical biosensor mice expressing such FRET-based biosensor molecules can provide 1) unperturbed physiological function, and 2) precise quantification of a molecular species, with high spatial and temporal resolution. FRET-based sensor molecules are available for studying a wide range of phenomena, from study of cell cycle to quantification of second messengers (e.g., cyclic nucleotides, inositol 1,4,5-trisphosphate, ions). Intramolecular FRET-based sensors, because they are ratiometric, are particularly important for in vivo cardiovascular research, where motion and blood flow are present. Measurement of a FRET ratio obviates problems related to contraction of arteries and to changes in blood flow, both of which can affect optical signals. A distinct advantage of the use of optical biosensor animals very much worth noting is that they facilitate the approach of observe, not perturb. Even a perfectly specific perturber (such as a specific Rho-kinase inhibitor) will cause disruption of multiple downstream pathways, with complicated results. Of course, the same is true even with highly specific inducible genetic interventions such as inducible knockdown of a particular RhoGEF.

I suggest that the present combination of genetic engineering and two-photon microscopy, realized in what we have called “optical biosensor mice,” makes possible a new level of in vivo experiments that can provide critical, previously inaccessible information. A key idea is to observe, not perturb, in the living animal, to probe the physiology or the disease state. Is it, to use that overused word a new paradigm? Not really, observation and measurement are the foundations of the scientific method. In reality, observation and measurement inevitably involve some perturbation, but careful construction of biosensor molecules and expression levels can minimize this effect.

Finally, in vivo experimentation, particularly of the type discussed here, represents a major investment of time and resources and involves new ethical concerns about use of animals. Nevertheless, I think a strong case for this new type of in vivo experimentation can be made, particularly for the study of hypertension. Billions of dollars have already been spent in hypertension research, and the result is largely the disparate identification of many individual genes, molecules, and organs that might be involved in hypertension. The fundamental cause and the mechanisms of hypertension remain unknown. We have argued recently that an integrated view of hypertension is required and is, in fact, emerging (2). In this view the central nervous system, the renal system, and the cardiovascular system are all dynamic partners in producing the hypertensive state. Thus an experimental approach now more viable than ever before is to determine what actually is altered in the living hypertensive animal. The ability to study conscious animals, for some types of experiments, may be particularly important, as arterial blood pressure is strongly affected by anesthesia. Only in the living animal does the nervous system exert dynamic control over artery diameter and renal function, and only in the living animal are arteries exposed to blood borne substances and the physical effects of blood flow, pressure and tissue metabolism. Yet, our quest for understanding is not satisfied until we know how the ongoing molecular events in these tissues are actually changed in the hypertensive state, not just those persisting in tissues removed from the animal or probed with loss or gain of function genetic alterations.

In summary, a compelling argument for rejuvenation of in vivo experimentation is that dynamic molecular events can now be observed in the living, and even conscious, animal, the only place where normal physiology and disease states really exist.


This work was supported by R01-HL-091969 (to W. G. Wier).


No conflicts of interest, financial or otherwise, are declared by the author.


W.G.W. drafted, edited and revised, and approved final version of manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.