Chapter 6Local control of ocular blood flow

6.6.1. Optic Nerve Head (ONH)

Numerous studies provide evidence of autoregulation in the prelaminar ONH in different species with blood flow measured with different techniques and perfusion pressure varied by changing IOP or blood pressure (Fig 6.15)^56–59. However, the underlying mechanisms responsible for this behavior are unclear, though metabolic and myogenic mechanisms seem most likely.

Several pieces of evidence indicate metabolic control in ONH autoregulation. First, the ONH has a strong functional hyperemic response during retinal illumination with a flickering light stimulus (Fig 6.16)⁶⁰. The functional hyperemia is linked to metabolism since it is associated with a decrease in ONH PO2 indicative of increased oxygen consumption and increased interstitial potassium concentration consistent with increased nerve activity⁶¹. Local NO production also increases and may contribute, although the functional hyperemia persists despite nitric oxide synthase inhibition⁶². The response is also altered by exogenous adenosine, but it is unclear whether endogenous adenosine is involved⁶². Second, the pressure-flow relationship is shifted upwards during metabolic stimulation (Fig 6.17)⁶³. Interestingly, the break-point in the pressure-flow relationship occurs at a higher perfusion pressure than in control. This may be due to fewer action potentials arriving from the retina due to insufficient retinal perfusion or the vasodilatory reserve may be exhausted sooner at the higher metabolic rate (i.e., the ONH arterioles have a maximum achievable diameter all of which is available to respond to decreased perfusion pressure under control conditions, but less is available when doubly challenged by increased metabolism and decreased perfusion pressure). Third, ONH blood flow decreases during hyperoxia and increases during hypoxia (Fig 6.18)⁶⁴. Lastly, the ONH appears to undergo a reactive hyperemia following brief periods of ischemia, at least in some species (Fig 6.19)⁶³. Reactive hyperemia seems to occur in cats, monkeys and humans⁶⁵, but the experiments were not designed specifically to study reactive hyperemia (i.e., the ischemia duration was not varied systematically or done at different levels of metabolic activity). Taken together, these lines of evidence support the involvement of metabolic local control in the ONH.

There is also evidence against metabolic involvement in ONH local control, at least in the rabbit (Fig 6.20). In that species, a 10 min period of relative ischemia did not elicit a reactive hyperemic response, although there was clear evidence of autoregulation (i.e., the relative decrease in ONH velocity was much smaller than the relative decrease in perfusion pressure)⁶⁸. Moreover, the speed of the autoregulatory response in the rabbit to a step-decrease in perfusion pressure was quite rapid (<5 sec), which seems fast for the accumulation of vasoactive metabolites and perhaps more in keeping with a myogenic response (though the flicker response, which is clearly metabolic, is also quite fast)⁶⁹. The difference in the reactive hyperemic responses in the rabbit versus the cat and monkey may reflect species differences, but more studies designed specifically to identify the local control mechanisms responsible for ONH autoregulation are clearly needed.

6.6.2. Choroid

The evidence regarding choroidal autoregulation has varied over the years. The advent of the microsphere technique for use in the eye⁷⁰ in the early 1970’s stimulated studies of choroidal responses to changing perfusion pressure, typically by raising IOP. In some cases, the results indicated no choroidal autoregulation, as in a cat study by Alm and Bill, while another cat study by Weiter et al. showed the choroid has autoregulatory ability (Fig 6.21)^71,72. In another study by Alm and Bill, the results in primates were ambiguous (Fig 6.22)⁷³.

Because they used femoral arterial pressure as an index of ophthalmic artery pressure, the authors noted “If the line connecting the values for the two eyes is extrapolated and then intercepts either the positive flow axis or the positive pressure axis below 25 cm H₂0 [18.4 mmHg] this suggests a vasodilation in the eye with reduced perfusion pressure.… Although some lines point towards the positive flow axis the mean result does not indicate any marked reduction in vascular resistance in response to a reduction in perfusion pressure.” Using the authors’ definition, roughly half the animals autoregulated and the other half did not (Fig 6.22). Given the lack of clear evidence of autoregulation, the authors concluded that the choroid was passive. Other studies came to a similar conclusion.

Clearer evidence of autoregulation in the choroid came later from rabbit studies using a fiber optic-based laser Doppler flowmeter (Fig 6.23)^17,74,75. In those studies, the efficacy of choroidal autoregulation depended on the method used to vary the perfusion pressure, i.e., autoregulation occurred over a wider perfusion pressure range when arterial pressure was manipulated without controlling IOP than when MAP was raised at a constant IOP or IOP was raised at a constant MAP. It was also noted that the efficacy of choroidal autoregulation depended on the method of anesthesia, i.e., autoregulation occurred under pentobarbital anesthesia but was largely abolished by Nembutal, which contains 40% propylene glycol and 10% alcohol⁷⁶. Methodological issues such as these and others may account for the discrepant evidence for choroidal autoregulation.

In contrast to the optic nerve and retina, the choroid is richly innervated, and so an interaction of local and neural control mechanisms is inherent in the choroidal response to changing perfusion pressure. Consequently, designating a choroidal blood flow response to changing perfusion pressure as autoregulation (i.e., due solely to local control) may not be appropriate. For that reason, some authors prefer the term “baroregulation” when all regulatory mechanisms are intact and choroidal blood flow does not change with perfusion pressure. However, in animals it is possible to minimize the neural control contribution by systemic ganglionic blockade (Fig 6.24). In rabbits, this causes a slight downward shift in the choroidal pressure-flow relationship, but choroidal blood flow still remains pressure-independent until perfusion pressure falls below approximately 40 mmHg⁷⁵. This result indicates that the choroid is capable of autoregulation. However, the downward shift in the pressure-flow relation with ganglionic blockade also indicates a degree of tonic neural vasodilator tone. Some of that tone may be nitridergic, since inhibition of nitric oxide synthase (NOS) with L-NAME also causes a downward shift in the pressure-flow relationship (Fig 6.24); however, L-NAME is a non-selective NOS inhibitor and much of the downward shift is likely due to the loss of endothelial NOS vasodilatory tone. The underlying vasoconstrictor tone revealed by NOS inhibition seems to be due to endothelin, since the L-NAME induced downward shift is largely reversed by the non-selective endothelin antagonist, A-182086^75,77.

Metabolic local control does not seem to play a role in choroidal autoregulation since the choroid does not respond markedly to retinal metabolic stimulation by flicker or to hyperoxia, though it does respond to hypocapnia (Fig 6.25)^78,79. Other studies report a hyperemic response to hypercapnia^71,80. The choroid also does not exhibit a reactive hyperemic response to brief (Fig 6.26) or long periods of ischemia⁸¹. Choroidal blood flow also does not increase (and may even decrease) when retinal oxygen consumption increases in the dark such that the inner photoreceptor layer becomes nearly anoxic (see below). The apparent absence of choroidal metabolic local control may be due to the high flow rate in the choroid and its anatomical organization. Metabolic local control is thought to depend on metabolic vasodilators reaching the resistance vessels. The choroid itself has little parenchymal metabolism, so the source of metabolic vasodilators would be the RPE and photoreceptors. The sheet-like choriocapillaris design is thought to optimize oxygen delivery and waste removal for the RPE and photoreceptors. If efficient, the waste removal function of the choriocapillaris would preclude vasodilators from the RPE and photoreceptors from reaching the choroidal arterioles. This is one plausible explanation for the lack of choroidal metabolic local control.

There is some evidence for choroidal myogenic local control, at least in the rabbit^24,74. To explore whether a myogenic mechanism could account for choroidal autoregulatory-like behavior in the rabbit, Kiel and Shepherd⁷⁴ created a mathematical model incorporating myogenic control of choroidal arterial resistance and passive (Starling resistor) control of choroidal venous resistance. The model simulated pressure-flow relationship was similar to that found in vivo, suggesting that a myogenic mechanism could account for the autoregulatory-like behavior. Moreover, the model predicted decreased choroidal blood flow in response to increased perfusion pressure in conditions conducive to heightened wall tension, and similar responses were found in vivo (Fig 6.27). Thus, it appears myogenic local control may be operative in the choroid.

As noted earlier, from a functional perspective, the myogenic mechanism is better suited to regulating capillary hydrostatic pressure than blood flow. If so, a myogenic mechanism in the choroid could help to stabilize blood volume during arterial pressure fluctuations and so minimize fluctuations in IOP. Consistent with this idea, the IOP normally rises a few mmHg in response to a 30–40 mmHg mechanically induced increase in arterial pressure, but if vascular control is blocked with a systemic vasodilator, the IOP response to a similar arterial pressure increase is much greater (Fig 6.28)²⁴. This IOP protective response occurs during autonomic blockade, so it has a local mechanism. If the arterial hypertension is neurally or humorally mediated, the local mechanism may augment the choroidal vasoconstriction relative to that occurring throughout the systemic circulation⁸².

A function proposed for the choroidal circulation that may have a local component is retinal temperature regulation. Given the small amount of metabolically active tissue (i.e., heat generating) relative to the size of the globe, the temperatures of the retina and the eye are primarily determined by the blood-borne convective heat delivery from the core and heat dissipation from the exposed ocular surface. As the largest of the ocular circulations, the choroid provides most of the convective heat delivery to the eye. In an environment cooler than core body temperature, the thermal gradient favors ocular heat loss; conversely, in a hot environment, the thermal gradient favors ocular heat gain. An additional source of heat gain is light absorption in the pigmented RPE and choroid. Given these thermodynamic parameters, it follows that eye temperature will fall if ocular blood flow is reduced in a cold environment, or rise if the environment is hot. If the temperature outside the eye is the same as core body temperature, stopping ocular blood flow will not change retinal temperature.

Parver et al. reported results from two monkeys in which raising IOP above systolic blood pressure decreased the temperature of a thermistor inside a 23 gauge needle inserted in the retina-choroid by 1° – 1.2° C⁸³. The temperatures of the environment, the saline drip on the cornea, and the infusate used to raise IOP were not provided so the thermal gradients are unclear. Given that the retina-choroid temperatures were measured at the macula where the thermal gradient should be dominated by the temperature of the orbit, the magnitude and speed (i.e., “not less than one minute after alteration of the pressure”) of the retina-choroid temperature decreases suggest that the thermal gradient to the exposed anterior eye surface must have been quite large. In the same two monkeys, exposure of the cornea to 1.09 mW/cm² of light caused an increase in retina-choroid temperature of ≈0.9 C when the IOP was set at 20 mmHg, which increased another ≈0.8 C when IOP was raised above systolic blood pressure.

Based on these results, the authors suggested “that the high flow choroidal circulation normally functions to stabilize the temperature environment of the retinal pigment epithelium and outer retinal layers.”⁸³ A problem with this suggestion is that the temperature of choroidal blood is set by the core body temperature, which can vary significantly. For example, walking for 30 min in a cool environment followed by walking for 1 hr in a hot environment can raise core body temperature by 3.5 C (Fig 6.29)84. Thus, in addition to the thermal gradient mentioned earlier, the ability of the choroid to stabilize retinal temperature depends greatly on whole body thermoregulation. Moreover, the 0.9 C increase in retina-choroid temperature upon light exposure with IOP at 20 mmHg, suggests that choroidal blood flow does a poor job of stabilizing retinal temperature. Indeed, some of that increase in temperature may have been a light-induced reflex since Parver et al. found that light applied to the contralateral eye increased ipsilateral retina-choroid and scleral temperature as well as an index of choroidal blood flow; the reflex, presumably, also works when the ipsilateral eye is light exposed⁸⁵. Such a reflex suggests that choroidal blood flow is not regulated to maintain retinal temperature.

6.6.3. Retina

Evidence for autoregulation of retinal blood flow is found in various species with different blood flow measuring techniques and methods of perfusion pressure manipulation (Fig 6.30). Given its lack of autonomic innervation and high metabolic needs, autoregulation of retinal blood flow is not surprising. However, local control in the retina is complex. In many species, retinal nutrient delivery and waste removal are provided by both retinal and choroidal circulations, but some species subsist with a negligible retinal circulation (e.g., rabbits), others have none (e.g., guinea pigs), and the primate foveal region is devoid of retinal vessels despite its high density of metabolically active photoreceptors. Obviously, the link between retinal perfusion and metabolism is complicated, and varies by species and location, which makes understanding metabolic local control difficult. The negative visual consequences of retinal edema underscore the likely importance of myogenic local control, but this mechanism is difficult to study in the in vivo retina though it is evident under in vitro conditions (Fig 6.1). The contributions of flow mediated vasodilation and intercellular conduction are even harder to study and less well understood. Thus, while the evidence for retinal metabolic local control predominates, the other forms of local control may contribute as well.

A counterintuitive phenomenon in the retina is that oxygen consumption rises in the dark due to increased Na/K ATPase activity^89,90. In species with a dual retinal blood supply (i.e., retinal and choroidal circulations), the dark stimulated increase in oxygen consumption is sufficient to lower the PO2 of the photoreceptor inner segments to near zero (Fig 6.31)^91,92. There is a corresponding increase in glucose consumption in approximately the same location in the dark, and retinal blood flow is also higher in the dark⁹³. This behavior appears to be an example of functional hyperemia, even though the metabolic activation occurs in the outer retina while the blood flow response occurs in the inner retina.

A clearer example of functional hyperemia is the retinal response to flickering light stimulation (Fig 6.32). In this case, the increase in retinal blood flow is associated with increased oxygen consumption indicated by the arteriovenous oxygen difference in paired retinal arteries and veins as well as increased glucose consumption in the inner retina^93–95. Interestingly, there appears to be the greatest increase in retinal blood flow in the area with the highest density of ganglion cells, consistent with a link between metabolic demand and perfusion⁹⁵. Additional evidence indicating significant retinal metabolic local control include the increase in retinal blood flow in response to hypoxia⁹⁶ and hypercapnia⁹⁷ as well as the decrease in blood flow in response to hyperoxia^98,99 (Fig 6.33) and the reactive hyperemia after brief (Fig 6.26) and long periods of ischemia (Fig 6.34)⁸¹. Surprisingly, the retinal pressure-flow relationships under light and dark-adapted conditions have not been determined, but it seems likely that the dark-adapted curve would be shifted upwards due to the heightened oxygen consumption. Similarly, reactive hyperemic responses to brief ischemic periods would be expected to be larger in the dark.

6.6.4. Ciliary body

While the retina has basically one function (converting light information for the central nervous system) and a dual blood supply, the ciliary body has two functions (accommodation and aqueous secretion) with a single blood supply. Differentiating ciliary muscle and secretory tissue is fairly simple by histology, but complicated in vivo, and measuring the blood flow to the two tissue types is even harder. Microsphere measurements at normal and elevated IOP indicate that ciliary muscle and ciliary processes both autoregulate, but the flow readings vary significantly, possibly because of the small number of spheres that can be trapped in such a small amount of tissue (Fig 6.35).

LDF measurements in the rabbit also provide evidence of ciliary autoregulation, although it is also clear that the ciliary circulation is under neural control, since ganglionic blockade caused an upward and leftward shift in the pressure-flow relationship indicating tonic neuroconstrictor tone (Fig 6.36). This is an interesting example of the interplay between neural and local control. Ciliary blood flow provides the oxygen and nutrient delivery to support aqueous production, and ciliary blood flow can decrease approximately 30% before aqueous production is compromised under control conditions. It may be this perfusion reserve that makes possible the constrained autoregulatory range seen when both neural and local control are operative. The local mechanisms involved in ciliary autoregulation have not been investigated systematically, but the author has rarely seen flow behavior suggestive of metabolic local control. For example, ciliary blood flow does not undergo a reactive hyperemic response following periods of ischemia (Fig 6.37).

6.6.5. Iris

There are only a few published studies of iris blood flow in animals or humans. Microsphere measurements in primates indicate the iris autoregulates, but the measurements were quite variable (Fig 6.38). Laser speckle flowgraphy measurements in pigmented rabbits suggest little iris autoregulatory ability (Fig 6.38). Human iris blood flow also does not appear to autoregulate, but oddly there is evidence of iridial reactive hyperemia¹⁰¹. Given the limited number of studies, definitive comments on iris local control require addition data.

Defining local control behavior and identifying the underlying mechanisms and their interactions is challenging even in more accessible vascular beds, but it is particularly hard in the ocular circulations. Local control clearly occurs in the eye, and further study will broaden our knowledge of its role in normal ocular function and disease.