Valency, Geometry, and Effective Avidity

Why Geometry Changes Avidity

When one arm of a bispecific antibody binds a target, its other arm is no longer exploring all of solution. It is tethered near the first interaction. If a second target lies within reach and is presented at a compatible orientation, the tethered arm can encounter it at a high effective local concentration. If spacing or orientation is unfavorable, that concentration can be very small—even when both arms bind their targets tightly in isolation.

Valency counts how many binding sites are available on the molecule; geometry determines how many of those sites can engage targets at once. Neither valency alone nor the intrinsic affinity of an arm specifies the probability of multivalent engagement. The calculations here condition on a first arm already being bound and ask how likely a second arm is to bind a nearby target. The earlier material on equilibrium avidity and kinetic rebinding provides related state and kinetic descriptions; the emphasis here is the geometric contribution to the local concentration.

Definition: The effective local concentration, $C_{\mathrm{eff}}$, is the concentration of a tethered binding arm that would give the same equilibrium binding odds to a nearby target as the arm’s geometric probability of reaching that target’s capture region. It is a conditional, geometry-dependent concentration, not the bulk concentration of antibody or target.

From Reachability to Effective Concentration

Suppose a first arm is bound and consider one candidate second target. Define $q$ as the probability that the tethered arm’s binding site is positioned and oriented within that target’s capture region. Let $V_{\mathrm{cap}}$ be the effective capture volume, in litres, associated with the target’s binding site. The local concentration that corresponds to this probability is

C_{\mathrm{eff}}=\frac{q}{N_A V_{\mathrm{cap}}},

where $N_A=6.022\times10^{23}\ \mathrm{mol}^{-1}$ is Avogadro’s constant. The product $N_A V_{\mathrm{cap}}$ has units of litres per mole, so $C_{\mathrm{eff}}$ has units of moles per litre. The probability $q$ combines positional reach and orientational compatibility. A target can be close in space but still have a low $q$ if the tethered paratope approaches from an inaccessible direction.

This expression is a useful coarse-grained representation, not a universal molecular-structure formula. A more detailed model may calculate the distribution of tethered-arm positions and orientations, then integrate over the target’s binding-compatible region. In either case, $q$ must refer to a defined molecular arrangement and a defined capture criterion. It is not simply the fraction of target molecules on a cell surface.

Equilibrium result: Let $K_d$ be the intrinsic dissociation constant of the second arm for its target, and assume the first arm remains bound. For one available second target, the ratio of doubly bound to singly bound states is $C_{\mathrm{eff}}/K_d$. Therefore, among these two states, $$ P(\text{second arm bound}\mid\text{first arm bound}) =\frac{C_{\mathrm{eff}}}{K_d+C_{\mathrm{eff}}}. $$

To justify the result, write the equilibrium odds for second-arm binding as the local concentration divided by its dissociation constant:

\frac{[D]}{[S]}=\frac{C_{\mathrm{eff}}}{K_d}, \qquad P(D\mid S\text{ or }D)=\frac{[D]}{[S]+[D]} =\frac{C_{\mathrm{eff}}/K_d}{1+C_{\mathrm{eff}}/K_d} =\frac{C_{\mathrm{eff}}}{K_d+C_{\mathrm{eff}}}.

Here $S$ denotes the singly engaged state and $D$ the state with the second arm bound. This conditional probability is not, by itself, an overall apparent $K_d$ for binding the bispecific to a cell or solution target. An overall apparent affinity also depends on how the first arm binds, target concentrations, the number of accessible target pairs, and the binding states included in the model.

Worked Comparisons

Worked Example: A Reachable Target Pair

Consider a hypothetical bispecific with a first arm already bound. Its second arm has intrinsic $K_d=0.50\ \mu\mathrm{M}$ for a nearby target. Suppose a geometry calculation gives $q=0.0060$, and the effective capture volume is $V_{\mathrm{cap}}=1.0\times10^{-21}\ \mathrm{L}$. Calculate the effective local concentration and conditional second-arm occupancy.

First calculate the conversion factor: $$ N_A V_{\mathrm{cap}} =(6.022\times10^{23}\ \mathrm{mol}^{-1})(1.0\times10^{-21}\ \mathrm{L}) =602.2\ \mathrm{L\,mol}^{-1}. $$

Then $$ C_{\mathrm{eff}}=\frac{0.0060}{602.2\ \mathrm{L\,mol}^{-1}} =9.96\times10^{-6}\ \mathrm{mol\,L}^{-1} =9.96\ \mu\mathrm{M}. $$ The conditional probability is $$ P(D\mid S\text{ or }D) =\frac{9.96}{9.96+0.50} =0.952. $$

Under the stated equilibrium assumptions, about 95.2% of the singly-or-doubly engaged population is doubly engaged. This is a prediction for this fixed geometry after first-arm binding, not a claim that 95.2% of all antibody molecules in a sample are bound.

Worked Example: Comparing Two Arm Arrangements

Two hypothetical bispecific formats have the same arm chemistry, the same $K_d=0.50\ \mu\mathrm{M}$, and the same capture volume $1.0\times10^{-21}\ \mathrm{L}$. Format A has a flexible arrangement for which $q=0.0060$. Format B has a more constrained arrangement for which $q=0.0012$. Compare their conditional second-arm engagement.

For Format A, the calculation above gives $C_{\mathrm{eff,A}}=9.96\ \mu\mathrm{M}$, so $$ P_A=\frac{9.96}{9.96+0.50}=0.952. $$ For Format B, $$ C_{\mathrm{eff,B}}=\frac{0.0012}{602.2\ \mathrm{L\,mol}^{-1}} =1.99\ \mu\mathrm{M}, \qquad P_B=\frac{1.99}{1.99+0.50}=0.799. $$

The fivefold difference in $q$ produces a fivefold difference in effective local concentration, but not a fivefold difference in occupancy: the occupancy relationship is nonlinear. In this hypothetical comparison, the flexible format has a conditional double-engagement probability of about 95.2%, versus about 79.9% for the constrained format. The comparison isolates geometry only because intrinsic affinity, capture volume, and the first-bound condition were held fixed.

Worked Example: More Than One Reachable Second Target

Suppose that after the first arm binds, a second arm can reach either of two distinct, non-overlapping capture regions. In a simplified calculation, let their conditional reach probabilities be $q_1=0.0020$ and $q_2=0.0030$, with $V_{\mathrm{cap}}=1.0\times10^{-21}\ \mathrm{L}$ for each. Assume both routes have the same intrinsic $K_d=0.50\ \mu\mathrm{M}$.

Because the capture regions are non-overlapping, their probabilities add: $$ q_{\mathrm{total}}=q_1+q_2=0.0050. $$ Thus $$ C_{\mathrm{eff,total}} =\frac{0.0050}{602.2\ \mathrm{L\,mol}^{-1}} =8.30\ \mu\mathrm{M}, \qquad P=\frac{8.30}{8.30+0.50}=0.943. $$

If only the second route were available, its effective concentration would be $$ C_{\mathrm{eff,2}}=\frac{0.0030}{602.2\ \mathrm{L\,mol}^{-1}} =4.98\ \mu\mathrm{M}, \qquad P_2=\frac{4.98}{4.98+0.50}=0.909. $$

The additional route raises the conditional probability from 90.9% to 94.3%. It does not double the probability, because occupancy approaches one as $C_{\mathrm{eff}}$ becomes large relative to $K_d$. The sum is valid here only because the routes are defined as non-overlapping possibilities. If the same configurations were counted in both $q_1$ and $q_2$, adding them would overcount reachability.

Valency, Geometry, and Model Choice

For a molecule with several possible second-arm routes, each route can be represented by its own effective concentration when the routes are mutually exclusive and share the same intrinsic $K_d$. Their contributions can then be summed to form a total effective concentration for that conditional step. If affinities differ, the state weights must instead retain each route separately; one should not combine them by adding probabilities without accounting for the corresponding $K_d$ values.

The key comparison between formats is therefore not simply “two arms versus one” or “more flexible versus more rigid.” It is whether the format’s arm arrangement produces accessible positions and orientations for the relevant target spacing. A longer reach may increase $q$ for a widely spaced pair while reducing orientational compatibility for a close pair. Flexibility can broaden the set of reachable separations, but it does not guarantee a higher local concentration at every separation: spreading configurations over a larger volume can reduce the probability density at any one location.

A geometry-based calculation is most useful as a transparent comparison under controlled assumptions. Alternatives include an explicit state-based equilibrium model, which represents each allowed bound state and its equilibrium weights, and a kinetic model, which tracks association, dissociation, and rebinding rates. The former is appropriate when multiple binding states and stoichiometries matter; the latter is needed when time-dependent engagement or dissociation is the question. A measured apparent affinity can summarize a particular assay, but it may combine geometry, target presentation, and assay conditions rather than identify their separate contributions.

Assumptions, Diagnostics, and Failure Modes

The conditional result above relies on a specific set of assumptions:

The first arm is bound, and its bound configuration is sufficiently stable for the second-step equilibrium to be considered.
The second arm has a defined intrinsic $K_d$ that applies in the geometric setting being modeled.
The target is available to the second arm, and the capture volume and reach probability describe the actual target spacing and orientation.
Binding is at equilibrium for the stated calculation; the result does not predict the time required to reach equilibrium.
The included singly and doubly bound states are the states relevant to the conditional probability. Additional states require an expanded model.

A prediction should be treated cautiously if it changes sharply with small, uncertain changes in assumed spacing, hinge flexibility, or capture volume. This sensitivity indicates that geometry is driving the answer and should be measured or explored over a plausible range. Other warning signs include predicted engagement when structural reach is impossible, summing overlapping capture routes, comparing formats with different arm affinities as if geometry were the only change, and interpreting a conditional occupancy as an overall binding fraction.

The coarse-grained $q$ also hides heterogeneity. A cell population may have a distribution of target spacings rather than one spacing, and a flexible antibody samples a distribution of conformations rather than one fixed reach. In such cases, calculate the conditional prediction across relevant geometries and average using the appropriate distribution; evaluating only a mean spacing can be misleading because the occupancy function is nonlinear. Likewise, if binding is not at equilibrium or the first arm frequently dissociates before the second engages, this equilibrium calculation alone cannot describe the observed time course.

Common pitfall: “The molecule has two arms, so its avidity must be twice the monovalent affinity.” Valency does not specify the probability of reaching a second target, and affinity is not generally multiplied by arm count. The correction is to define the first-bound condition, estimate each accessible route’s $C_{\mathrm{eff}}$, and compare that value with the intrinsic $K_d$, while keeping conditional engagement distinct from overall apparent affinity.

Practical Interpretation

A useful report of a geometry-based avidity prediction states the assumed first-bound configuration, target spacing and orientation, intrinsic arm $K_d$, capture-volume convention, and how $q$ was obtained. It should report $C_{\mathrm{eff}}$ with concentration units and the resulting conditional second-arm probability. If geometry is uncertain, report a range rather than a single precise value. These details make it possible to distinguish a genuine format effect from an effect caused by different target presentation or modeling assumptions.

The central conclusion is conditional but useful: when $C_{\mathrm{eff}}$ is much greater than $K_d$, second-arm engagement is likely once the first arm has bound; when it is much less than $K_d$, second engagement is unlikely. Between those regimes, modest changes in spacing, orientation, or accessible routes can meaningfully change predicted avidity. The calculation organizes those effects without mistaking valency for geometry or conditional avidity for a universal affinity constant.

Check Your Understanding

Answer each question, then compare your reasoning with the solutions.

If $C_{\mathrm{eff}}=2.0\ \mu\mathrm{M}$ and $K_d=0.50\ \mu\mathrm{M}$, what is the conditional probability of second-arm binding?
Two formats have the same intrinsic $K_d$ and capture volume, but one has twice the reach probability $q$. Is its conditional occupancy necessarily twice as large?
Why can two non-overlapping second-target routes contribute additively to effective concentration in the simplified model, and when would that addition be invalid?
A model predicts high conditional second-arm engagement. Does that establish that most antibody molecules in the assay are doubly bound? Explain.

Solutions. (1) $P=2.0/(2.0+0.50)=0.800$, or 80.0%. (2) No. Doubling $q$ doubles $C_{\mathrm{eff}}$ under fixed $N_A V_{\mathrm{cap}}$, but occupancy is $C_{\mathrm{eff}}/(K_d+C_{\mathrm{eff}})$, a nonlinear function. (3) Mutually exclusive, non-overlapping capture regions represent distinct reachable configurations, so their probabilities can be summed. If the routes overlap, the shared configurations are counted twice and must be removed. (4) No. The result is conditional on the first arm already being bound and concerns the modeled singly and doubly engaged states; the overall fraction also depends on first-arm binding, target availability, and other states.